-#include "float.h"
-
-/* evaluate RISCV DSP feature */
-#if (defined (__RISCV_FEATURE_DSP) && (__RISCV_FEATURE_DSP == 1))
- #define RISCV_MATH_DSP
-#endif
-
- /**
- * @brief Macros required for reciprocal calculation in Normalized LMS
- */
-
-#define DELTA_Q31 (0x100)
-#define DELTA_Q15 0x5
-#define INDEX_MASK 0x0000003F
-#ifndef PI
- #define PI 3.14159265358979f
-#endif
-
- /**
- * @brief Macros required for SINE and COSINE Fast math approximations
- */
-
-#define FAST_MATH_TABLE_SIZE 512
-#define FAST_MATH_Q31_SHIFT (32 - 10)
-#define FAST_MATH_Q15_SHIFT (16 - 10)
-#define CONTROLLER_Q31_SHIFT (32 - 9)
-#define TABLE_SPACING_Q31 0x400000
-#define TABLE_SPACING_Q15 0x80
-
- /**
- * @brief Macros required for SINE and COSINE Controller functions
- */
- /* 1.31(q31) Fixed value of 2/360 */
- /* -1 to +1 is divided into 360 values so total spacing is (2/360) */
-#define INPUT_SPACING 0xB60B61
-
-
- /**
- * @brief Error status returned by some functions in the library.
- */
-
- typedef enum
- {
- RISCV_MATH_SUCCESS = 0, /**< No error */
- RISCV_MATH_ARGUMENT_ERROR = -1, /**< One or more arguments are incorrect */
- RISCV_MATH_LENGTH_ERROR = -2, /**< Length of data buffer is incorrect */
- RISCV_MATH_SIZE_MISMATCH = -3, /**< Size of matrices is not compatible with the operation */
- RISCV_MATH_NANINF = -4, /**< Not-a-number (NaN) or infinity is generated */
- RISCV_MATH_SINGULAR = -5, /**< Input matrix is singular and cannot be inverted */
- RISCV_MATH_TEST_FAILURE = -6 /**< Test Failed */
- } riscv_status;
-
- /**
- * @brief 8-bit fractional data type in 1.7 format.
- */
- typedef int8_t q7_t;
-
- /**
- * @brief 16-bit fractional data type in 1.15 format.
- */
- typedef int16_t q15_t;
-
- /**
- * @brief 32-bit fractional data type in 1.31 format.
- */
- typedef int32_t q31_t;
-
- /**
- * @brief 64-bit fractional data type in 1.63 format.
- */
- typedef int64_t q63_t;
-
- /**
- * @brief 32-bit floating-point type definition.
- */
- typedef float float32_t;
-
- /**
- * @brief 64-bit floating-point type definition.
- */
- typedef double float64_t;
-
-
-/**
- @brief definition to read/write two 16 bit values.
- @deprecated
- */
-#define __SIMD32_TYPE int32_t
-
-#define __SIMD32(addr) (*(__SIMD32_TYPE **) & (addr))
-#define __SIMD32_CONST(addr) ( (__SIMD32_TYPE * ) (addr))
-#define _SIMD32_OFFSET(addr) (*(__SIMD32_TYPE * ) (addr))
-#define __SIMD64(addr) (*( int64_t **) & (addr))
-
-/* SIMD replacement */
-
-/**
- @brief Read 2 Q31 from Q31 pointer and increment pointer afterwards.
- @param[in] pQ31 points to input value
- @return Q63 value
- */
-__STATIC_FORCEINLINE q63_t read_q31x2_ia (
- q31_t ** pQ31)
-{
- q63_t val;
-#ifndef RISCV_ALIGN_ACCESS
-#if __RISCV_XLEN == 64
- val = __LD(*pQ31);
-#else
- val = *((q63_t *)*pQ31);
-#endif /* __RISCV_XLEN == 64 */
-#else
- memcpy((void *)(&val), (void *)(*pQ31), 8);
-#endif
- *pQ31 += 2;
- return (val);
-}
-
-/**
- @brief Read 2 Q31 from Q31 pointer and decrement pointer afterwards.
- @param[in] pQ31 points to input value
- @return Q63 value
- */
-__STATIC_FORCEINLINE q63_t read_q31x2_da (
- q31_t ** pQ31)
-{
- q63_t val;
-#ifndef RISCV_ALIGN_ACCESS
-#if __RISCV_XLEN == 64
- val = __LD(*pQ31);
-#else
- val = *((q63_t *)*pQ31);
-#endif /* __RISCV_XLEN == 64 */
-#else
- memcpy((void *)(&val), (void *)(*pQ31), 8);
-#endif
- *pQ31 -= 2;
- return (val);
-}
-
-/**
- @brief Read 2 Q31 from Q31 pointer.
- @param[in] pQ31 points to input value
- @return Q63 value
- */
-__STATIC_FORCEINLINE q63_t read_q31x2 (
- q31_t * pQ31)
-{
- q63_t val;
-#ifndef RISCV_ALIGN_ACCESS
-#if __RISCV_XLEN == 64
- val = __LD(pQ31);
-#else
- val = *((q63_t *)pQ31);
-#endif /* __RISCV_XLEN == 64 */
-#else
- memcpy((void *)(&val), (void *)(pQ31), 8);
-#endif
- return (val);
-}
-
-/**
- @brief Write 2 Q31 to Q31 pointer and increment pointer afterwards.
- @param[in] pQ31 points to input value
- @param[in] value Q63 value
- @return none
- */
-__STATIC_FORCEINLINE void write_q31x2_ia (
- q31_t ** pQ31,
- q63_t value)
-{
-#ifndef RISCV_ALIGN_ACCESS
-#if __RISCV_XLEN == 64
- __SD(*pQ31, value);
-#else
- *((q63_t *)*pQ31) = value;
-#endif /* __RISCV_XLEN == 64 */
-#else
- memcpy((void *)(*pQ31), (void *)(&value), 8);
-#endif
- *pQ31 += 2;
-}
-
-/**
- @brief Write 2 Q31 to Q31 pointer.
- @param[in] pQ31 points to input value
- @param[in] value Q63 value
- @return none
- */
-__STATIC_FORCEINLINE void write_q31x2 (
- q31_t * pQ31,
- q63_t value)
-{
-#ifndef RISCV_ALIGN_ACCESS
-#if __RISCV_XLEN == 64
- __SD(pQ31, value);
-#else
- *((q63_t *)pQ31) = value;
-#endif /* __RISCV_XLEN == 64 */
-#else
- memcpy((void *)(pQ31), (void *)(&value), 8);
-#endif
-}
-
-/**
- @brief Read 2 Q15 from Q15 pointer.
- @param[in] pQ15 points to input value
- @return Q31 value
- */
-__STATIC_FORCEINLINE q31_t read_q15x2 (
- q15_t * pQ15)
-{
- q31_t val;
-
-#ifndef RISCV_ALIGN_ACCESS
- __ASM volatile (
- "lw %0, (%1)"
- :"=r"(val)
- :"r"(pQ15)
- );
-#else
- memcpy((void *)(&val), (void *)(pQ15), 4);
-#endif
- return (val);
-}
-
-/**
- @brief Read 2 Q15 from Q15 pointer and increment pointer afterwards.
- @param[in] pQ15 points to input value
- @return Q31 value
- */
-__STATIC_FORCEINLINE q31_t read_q15x2_ia (
- q15_t ** pQ15)
-{
- q31_t val;
-
-#ifndef RISCV_ALIGN_ACCESS
- __ASM volatile (
- "lw %0, (%1)"
- :"=r"(val)
- :"r"(*pQ15)
- );
-#else
- memcpy((void *)(&val), (void *)(*pQ15), 4);
-#endif
- *pQ15 += 2;
-
- return (val);
-}
-
-/**
- @brief Read 4 Q15 from Q15 pointer and increment pointer afterwards.
- @param[in] pQ15 points to input value
- @return Q63 value
- */
-__STATIC_FORCEINLINE q63_t read_q15x4_ia (
- q15_t ** pQ15)
-{
- q63_t val;
-#ifndef RISCV_ALIGN_ACCESS
- val = *((q63_t *)*pQ15);
-#else
- memcpy((void *)(&val), (void *)(*pQ15), 8);
-#endif
- *pQ15 += 4;
-
- return (val);
-}
-
-/**
- @brief Read 4 Q15 from Q15 pointer.
- @param[in] pQ15 points to input value
- @return Q63 value
- */
-__STATIC_FORCEINLINE q63_t read_q15x4 (
- q15_t * pQ15)
-{
- q63_t val;
-#ifndef RISCV_ALIGN_ACCESS
-#if __RISCV_XLEN == 64
- val = __LD(pQ15);
-#else
- val = *((q63_t *)pQ15);
-#endif /* __RISCV_XLEN == 64 */
-#else
- memcpy((void *)(&val), (void *)(pQ15), 8);
-#endif
- return (val);
-}
-
-/**
- @brief Read 2 Q15 from Q15 pointer and decrement pointer afterwards.
- @param[in] pQ15 points to input value
- @return Q31 value
- */
-__STATIC_FORCEINLINE q31_t read_q15x2_da (
- q15_t ** pQ15)
-{
- q31_t val;
-
-#ifndef RISCV_ALIGN_ACCESS
- __ASM volatile (
- "lw %0, (%1)"
- :"=r"(val)
- :"r"(*pQ15)
- );
-#else
- memcpy((void *)(&val), (void *)(*pQ15), 4);
-#endif
- *pQ15 -= 2;
-
- return (val);
-}
-
-/**
- @brief Read 4 Q15 from Q15 pointer and decrement pointer afterwards.
- @param[in] pQ15 points to input value
- @return Q31 value
- */
-__STATIC_FORCEINLINE q63_t read_q15x4_da (
- q15_t ** pQ15)
-{
- q63_t val;
-#ifndef RISCV_ALIGN_ACCESS
- val = *((q63_t *)*pQ15);
-#else
- memcpy((void *)(&val), (void *)(*pQ15), 8);
-#endif
- *pQ15 -= 4;
-
- return (val);
-}
-
-/**
- @brief Write 2 Q15 to Q15 pointer and increment pointer afterwards.
- @param[in] pQ15 points to input value
- @param[in] value Q31 value
- @return none
- */
-__STATIC_FORCEINLINE void write_q15x2_ia (
- q15_t ** pQ15,
- q31_t value)
-{
-#ifndef RISCV_ALIGN_ACCESS
- __ASM volatile (
- "sw %0, (%1)"
- :
- :"r"(value), "r"(*pQ15)
- :"memory"
- );
-#else
- memcpy((void *)(*pQ15), (void *)(&value), 4);
-#endif
- *pQ15 += 2;
-}
-
-/**
- @brief Write 4 Q15 to Q15 pointer and increment pointer afterwards.
- @param[in] pQ15 points to input value
- @param[in] value Q31 value
- @return none
- */
-__STATIC_FORCEINLINE void write_q15x4_ia (
- q15_t ** pQ15,
- q63_t value)
-{
-#ifndef RISCV_ALIGN_ACCESS
- *((q63_t *)*pQ15) = value;
-#else
- memcpy((void *)(*pQ15), (void *)(&value), 8);
-#endif
- *pQ15 += 4;
-}
-
-/**
- @brief Write 4 Q15 to Q15 pointer and decrement pointer afterwards.
- @param[in] pQ15 points to input value
- @param[in] value Q31 value
- @return none
- */
-__STATIC_FORCEINLINE void write_q15x4_da (
- q15_t ** pQ15,
- q63_t value)
-{
-#ifndef RISCV_ALIGN_ACCESS
- *((q63_t *)*pQ15) = value;
-#else
- memcpy((void *)(*pQ15), (void *)(&value), 8);
-#endif
- *pQ15 -= 4;
-}
-
-/**
- @brief Write 2 Q15 to Q15 pointer.
- @param[in] pQ15 points to input value
- @param[in] value Q31 value
- @return none
- */
-__STATIC_FORCEINLINE void write_q15x2 (
- q15_t * pQ15,
- q31_t value)
-{
-#ifndef RISCV_ALIGN_ACCESS
- __ASM volatile (
- "sw %0, (%1)"
- :
- :"r"(value), "r"(pQ15)
- :"memory"
- );
-#else
- memcpy((void *)(pQ15), (void *)(&value), 4);
-#endif
-
-}
-
-/**
- @brief Write 4 Q15 to Q15 pointer.
- @param[in] pQ15 points to input value
- @param[in] value Q31 value
- @return none
- */
-__STATIC_FORCEINLINE void write_q15x4 (
- q15_t * pQ15,
- q63_t value)
-{
-#ifndef RISCV_ALIGN_ACCESS
- *((q63_t *)pQ15) = value;
-#else
- memcpy((void *)(pQ15), (void *)(&value), 8);
-#endif
-}
-
-/**
- @brief Read 8 Q7 from Q7 pointer and increment pointer afterwards.
- @param[in] pQ7 points to input value
- @return Q63 value
- */
-__STATIC_FORCEINLINE q63_t read_q7x8_ia (
- q7_t ** pQ7)
-{
- q63_t val;
-#ifndef RISCV_ALIGN_ACCESS
- val = *((q63_t *)*pQ7);
-#else
- memcpy((void *)(&val), (void *)(*pQ7), 8);
-#endif
- *pQ7 += 8;
-
- return val;
-}
-
-/**
- @brief Read 8 Q7 from Q7 pointer and decrement pointer afterwards.
- @param[in] pQ7 points to input value
- @return Q63 value
- */
-__STATIC_FORCEINLINE q63_t read_q7x8_da (
- q7_t ** pQ7)
-{
- q63_t val;
-#ifndef RISCV_ALIGN_ACCESS
- val = *((q63_t *)*pQ7);
-#else
- memcpy((void *)(&val), (void *)(*pQ7), 8);
-#endif
- *pQ7 -= 8;
- return val;
-}
-
-/**
- @brief Read 4 Q7 from Q7 pointer and increment pointer afterwards.
- @param[in] pQ7 points to input value
- @return Q31 value
- */
-__STATIC_FORCEINLINE q31_t read_q7x4_ia (
- q7_t ** pQ7)
-{
- q31_t val;
-
-#ifndef RISCV_ALIGN_ACCESS
- __ASM volatile (
- "lw %0, (%1)"
- :"=r"(val)
- :"r"(*pQ7)
- );
-#else
- memcpy((void *)(&val), (void *)(*pQ7), 4);
-#endif
- *pQ7 += 4;
-
- return (val);
-}
-
-/**
- @brief Read 4 Q7 from Q7 pointer and decrement pointer afterwards.
- @param[in] pQ7 points to input value
- @return Q31 value
- */
-__STATIC_FORCEINLINE q31_t read_q7x4_da (
- q7_t ** pQ7)
-{
- q31_t val;
-
-#ifndef RISCV_ALIGN_ACCESS
- __ASM volatile (
- "lw %0, (%1)"
- :"=r"(val)
- :"r"(*pQ7)
- );
-#else
- memcpy((void *)(&val), (void *)(*pQ7), 4);
-#endif
- *pQ7 -= 4;
-
- return (val);
-}
-
-/**
- @brief Write 8 Q7 to Q7 pointer and increment pointer afterwards.
- @param[in] pQ7 points to input value
- @param[in] value Q63 value
- @return none
- */
-__STATIC_FORCEINLINE void write_q7x8_ia (
- q7_t ** pQ7,
- q63_t value)
-{
-#ifndef RISCV_ALIGN_ACCESS
- *((q63_t *)*pQ7) = value;
-#else
- memcpy((void *)(*pQ7), (void *)(&value), 8);
-#endif
- *pQ7 += 8;
-}
-
-/**
- @brief Write 4 Q7 to Q7 pointer and increment pointer afterwards.
- @param[in] pQ7 points to input value
- @param[in] value Q31 value
- @return none
- */
-__STATIC_FORCEINLINE void write_q7x4_ia (
- q7_t ** pQ7,
- q31_t value)
-{
- q31_t val = value;
-
-#ifndef RISCV_ALIGN_ACCESS
- __ASM volatile (
- "sw %0, (%1)"
- :
- :"r"(value), "r"(*pQ7)
- :"memory"
- );
-#else
- memcpy((void *)(*pQ7), (void *)(&value), 4);
-#endif
- *pQ7 += 4;
-}
-
-/**
-* @brief definition to pack four 8 bit values.
-*/
-#define __PACKq7(v0,v1,v2,v3) ( (((int32_t)(v0) << 0) & (int32_t)0x000000FF) | \
- (((int32_t)(v1) << 8) & (int32_t)0x0000FF00) | \
- (((int32_t)(v2) << 16) & (int32_t)0x00FF0000) | \
- (((int32_t)(v3) << 24) & (int32_t)0xFF000000) )
-
-
-
- /**
- * @brief Clips Q63 to Q31 values.
- */
- __STATIC_FORCEINLINE q31_t clip_q63_to_q31(
- q63_t x)
- {
- return ((q31_t) (x >> 32) != ((q31_t) x >> 31)) ?
- ((0x7FFFFFFF ^ ((q31_t) (x >> 63)))) : (q31_t) x;
- }
-
- /**
- * @brief Clips Q63 to Q15 values.
- */
- __STATIC_FORCEINLINE q15_t clip_q63_to_q15(
- q63_t x)
- {
- return ((q31_t) (x >> 32) != ((q31_t) x >> 31)) ?
- ((0x7FFF ^ ((q15_t) (x >> 63)))) : (q15_t) (x >> 15);
- }
-
- /**
- * @brief Clips Q31 to Q7 values.
- */
- __STATIC_FORCEINLINE q7_t clip_q31_to_q7(
- q31_t x)
- {
- return ((q31_t) (x >> 24) != ((q31_t) x >> 23)) ?
- ((0x7F ^ ((q7_t) (x >> 31)))) : (q7_t) x;
- }
-
- /**
- * @brief Clips Q31 to Q15 values.
- */
- __STATIC_FORCEINLINE q15_t clip_q31_to_q15(
- q31_t x)
- {
- return ((q31_t) (x >> 16) != ((q31_t) x >> 15)) ?
- ((0x7FFF ^ ((q15_t) (x >> 31)))) : (q15_t) x;
- }
-
- /**
- * @brief Multiplies 32 X 64 and returns 32 bit result in 2.30 format.
- */
- __STATIC_FORCEINLINE q63_t mult32x64(
- q63_t x,
- q31_t y)
- {
- return ((((q63_t) (x & 0x00000000FFFFFFFF) * y) >> 32) +
- (((q63_t) (x >> 32) * y) ) );
- }
-
- /**
- * @brief Function to Calculates 1/in (reciprocal) value of Q31 Data type.
- */
- __STATIC_FORCEINLINE uint32_t riscv_recip_q31(
- q31_t in,
- q31_t * dst,
- const q31_t * pRecipTable)
- {
- q31_t out;
- uint32_t tempVal;
- uint32_t index, i;
- uint32_t signBits;
-
- if (in > 0)
- {
- signBits = ((uint32_t) (__CLZ( in) - 1));
- }
- else
- {
- signBits = ((uint32_t) (__CLZ(-in) - 1));
- }
-
- /* Convert input sample to 1.31 format */
- in = (in << signBits);
-
- /* calculation of index for initial approximated Val */
- index = (uint32_t)(in >> 24);
- index = (index & INDEX_MASK);
-
- /* 1.31 with exp 1 */
- out = pRecipTable[index];
-
- /* calculation of reciprocal value */
- /* running approximation for two iterations */
- for (i = 0U; i < 2U; i++)
- {
- tempVal = (uint32_t) (((q63_t) in * out) >> 31);
- tempVal = 0x7FFFFFFFu - tempVal;
- /* 1.31 with exp 1 */
- /* out = (q31_t) (((q63_t) out * tempVal) >> 30); */
- out = clip_q63_to_q31(((q63_t) out * tempVal) >> 30);
- }
-
- /* write output */
- *dst = out;
-
- /* return num of signbits of out = 1/in value */
- return (signBits + 1U);
- }
-
-
- /**
- * @brief Function to Calculates 1/in (reciprocal) value of Q15 Data type.
- */
- __STATIC_FORCEINLINE uint32_t riscv_recip_q15(
- q15_t in,
- q15_t * dst,
- const q15_t * pRecipTable)
- {
- q15_t out = 0;
- uint32_t tempVal = 0;
- uint32_t index = 0, i = 0;
- uint32_t signBits = 0;
-
- if (in > 0)
- {
- signBits = ((uint32_t)(__CLZ( in) - 17));
- }
- else
- {
- signBits = ((uint32_t)(__CLZ(-in) - 17));
- }
-
- /* Convert input sample to 1.15 format */
- in = (in << signBits);
-
- /* calculation of index for initial approximated Val */
- index = (uint32_t)(in >> 8);
- index = (index & INDEX_MASK);
-
- /* 1.15 with exp 1 */
- out = pRecipTable[index];
-
- /* calculation of reciprocal value */
- /* running approximation for two iterations */
- for (i = 0U; i < 2U; i++)
- {
- tempVal = (uint32_t) (((q31_t) in * out) >> 15);
- tempVal = 0x7FFFu - tempVal;
- /* 1.15 with exp 1 */
- out = (q15_t) (((q31_t) out * tempVal) >> 14);
- /* out = clip_q31_to_q15(((q31_t) out * tempVal) >> 14); */
- }
-
- /* write output */
- *dst = out;
-
- /* return num of signbits of out = 1/in value */
- return (signBits + 1);
- }
-
-
-/*
- * @brief C custom defined intrinsic functions
- */
-#if !defined (RISCV_MATH_DSP)
-
- /*
- * @brief C custom defined QADD8
- */
- __STATIC_FORCEINLINE uint32_t __QADD8(
- uint32_t x,
- uint32_t y)
- {
- q31_t r, s, t, u;
-
- r = __SSAT(((((q31_t)x << 24) >> 24) + (((q31_t)y << 24) >> 24)), 8) & (int32_t)0x000000FF;
- s = __SSAT(((((q31_t)x << 16) >> 24) + (((q31_t)y << 16) >> 24)), 8) & (int32_t)0x000000FF;
- t = __SSAT(((((q31_t)x << 8) >> 24) + (((q31_t)y << 8) >> 24)), 8) & (int32_t)0x000000FF;
- u = __SSAT(((((q31_t)x ) >> 24) + (((q31_t)y ) >> 24)), 8) & (int32_t)0x000000FF;
-
- return ((uint32_t)((u << 24) | (t << 16) | (s << 8) | (r )));
- }
-
-
- /*
- * @brief C custom defined QSUB8
- */
- __STATIC_FORCEINLINE uint32_t __QSUB8(
- uint32_t x,
- uint32_t y)
- {
- q31_t r, s, t, u;
-
- r = __SSAT(((((q31_t)x << 24) >> 24) - (((q31_t)y << 24) >> 24)), 8) & (int32_t)0x000000FF;
- s = __SSAT(((((q31_t)x << 16) >> 24) - (((q31_t)y << 16) >> 24)), 8) & (int32_t)0x000000FF;
- t = __SSAT(((((q31_t)x << 8) >> 24) - (((q31_t)y << 8) >> 24)), 8) & (int32_t)0x000000FF;
- u = __SSAT(((((q31_t)x ) >> 24) - (((q31_t)y ) >> 24)), 8) & (int32_t)0x000000FF;
-
- return ((uint32_t)((u << 24) | (t << 16) | (s << 8) | (r )));
- }
-
-
- /*
- * @brief C custom defined QADD16
- */
- __STATIC_FORCEINLINE uint32_t __QADD16(
- uint32_t x,
- uint32_t y)
- {
-/* q31_t r, s; without initialisation 'riscv_offset_q15 test' fails but 'intrinsic' tests pass! for armCC */
- q31_t r = 0, s = 0;
-
- r = __SSAT(((((q31_t)x << 16) >> 16) + (((q31_t)y << 16) >> 16)), 16) & (int32_t)0x0000FFFF;
- s = __SSAT(((((q31_t)x ) >> 16) + (((q31_t)y ) >> 16)), 16) & (int32_t)0x0000FFFF;
-
- return ((uint32_t)((s << 16) | (r )));
- }
-
-
- /*
- * @brief C custom defined SHADD16
- */
- __STATIC_FORCEINLINE uint32_t __SHADD16(
- uint32_t x,
- uint32_t y)
- {
- q31_t r, s;
-
- r = (((((q31_t)x << 16) >> 16) + (((q31_t)y << 16) >> 16)) >> 1) & (int32_t)0x0000FFFF;
- s = (((((q31_t)x ) >> 16) + (((q31_t)y ) >> 16)) >> 1) & (int32_t)0x0000FFFF;
-
- return ((uint32_t)((s << 16) | (r )));
- }
-
-
- /*
- * @brief C custom defined QSUB16
- */
- __STATIC_FORCEINLINE uint32_t __QSUB16(
- uint32_t x,
- uint32_t y)
- {
- q31_t r, s;
-
- r = __SSAT(((((q31_t)x << 16) >> 16) - (((q31_t)y << 16) >> 16)), 16) & (int32_t)0x0000FFFF;
- s = __SSAT(((((q31_t)x ) >> 16) - (((q31_t)y ) >> 16)), 16) & (int32_t)0x0000FFFF;
-
- return ((uint32_t)((s << 16) | (r )));
- }
-
-
- /*
- * @brief C custom defined SHSUB16
- */
- __STATIC_FORCEINLINE uint32_t __SHSUB16(
- uint32_t x,
- uint32_t y)
- {
- q31_t r, s;
-
- r = (((((q31_t)x << 16) >> 16) - (((q31_t)y << 16) >> 16)) >> 1) & (int32_t)0x0000FFFF;
- s = (((((q31_t)x ) >> 16) - (((q31_t)y ) >> 16)) >> 1) & (int32_t)0x0000FFFF;
-
- return ((uint32_t)((s << 16) | (r )));
- }
-
-
- /*
- * @brief C custom defined QASX
- */
- __STATIC_FORCEINLINE uint32_t __QASX(
- uint32_t x,
- uint32_t y)
- {
- q31_t r, s;
-
- r = __SSAT(((((q31_t)x << 16) >> 16) - (((q31_t)y ) >> 16)), 16) & (int32_t)0x0000FFFF;
- s = __SSAT(((((q31_t)x ) >> 16) + (((q31_t)y << 16) >> 16)), 16) & (int32_t)0x0000FFFF;
-
- return ((uint32_t)((s << 16) | (r )));
- }
-
-
- /*
- * @brief C custom defined SHASX
- */
- __STATIC_FORCEINLINE uint32_t __SHASX(
- uint32_t x,
- uint32_t y)
- {
- q31_t r, s;
-
- r = (((((q31_t)x << 16) >> 16) - (((q31_t)y ) >> 16)) >> 1) & (int32_t)0x0000FFFF;
- s = (((((q31_t)x ) >> 16) + (((q31_t)y << 16) >> 16)) >> 1) & (int32_t)0x0000FFFF;
-
- return ((uint32_t)((s << 16) | (r )));
- }
-
-
- /*
- * @brief C custom defined QSAX
- */
- __STATIC_FORCEINLINE uint32_t __QSAX(
- uint32_t x,
- uint32_t y)
- {
- q31_t r, s;
-
- r = __SSAT(((((q31_t)x << 16) >> 16) + (((q31_t)y ) >> 16)), 16) & (int32_t)0x0000FFFF;
- s = __SSAT(((((q31_t)x ) >> 16) - (((q31_t)y << 16) >> 16)), 16) & (int32_t)0x0000FFFF;
-
- return ((uint32_t)((s << 16) | (r )));
- }
-
-
- /*
- * @brief C custom defined SHSAX
- */
- __STATIC_FORCEINLINE uint32_t __SHSAX(
- uint32_t x,
- uint32_t y)
- {
- q31_t r, s;
-
- r = (((((q31_t)x << 16) >> 16) + (((q31_t)y ) >> 16)) >> 1) & (int32_t)0x0000FFFF;
- s = (((((q31_t)x ) >> 16) - (((q31_t)y << 16) >> 16)) >> 1) & (int32_t)0x0000FFFF;
-
- return ((uint32_t)((s << 16) | (r )));
- }
-
-
- /*
- * @brief C custom defined SMUSDX
- */
- __STATIC_FORCEINLINE uint32_t __SMUSDX(
- uint32_t x,
- uint32_t y)
- {
- return ((uint32_t)(((((q31_t)x << 16) >> 16) * (((q31_t)y ) >> 16)) -
- ((((q31_t)x ) >> 16) * (((q31_t)y << 16) >> 16)) ));
- }
-
- /*
- * @brief C custom defined SMUADX
- */
- __STATIC_FORCEINLINE uint32_t __SMUADX(
- uint32_t x,
- uint32_t y)
- {
- return ((uint32_t)(((((q31_t)x << 16) >> 16) * (((q31_t)y ) >> 16)) +
- ((((q31_t)x ) >> 16) * (((q31_t)y << 16) >> 16)) ));
- }
-
-
- /*
- * @brief C custom defined QADD
- */
- __STATIC_FORCEINLINE int32_t __QADD(
- int32_t x,
- int32_t y)
- {
- return ((int32_t)(clip_q63_to_q31((q63_t)x + (q31_t)y)));
- }
-
-
- /*
- * @brief C custom defined QSUB
- */
- __STATIC_FORCEINLINE int32_t __QSUB(
- int32_t x,
- int32_t y)
- {
- return ((int32_t)(clip_q63_to_q31((q63_t)x - (q31_t)y)));
- }
-
-
- /*
- * @brief C custom defined SMLAD
- */
- __STATIC_FORCEINLINE uint32_t __SMLAD(
- uint32_t x,
- uint32_t y,
- uint32_t sum)
- {
- return ((uint32_t)(((((q31_t)x << 16) >> 16) * (((q31_t)y << 16) >> 16)) +
- ((((q31_t)x ) >> 16) * (((q31_t)y ) >> 16)) +
- ( ((q31_t)sum ) ) ));
- }
-
-
- /*
- * @brief C custom defined SMLADX
- */
- __STATIC_FORCEINLINE uint32_t __SMLADX(
- uint32_t x,
- uint32_t y,
- uint32_t sum)
- {
- return ((uint32_t)(((((q31_t)x << 16) >> 16) * (((q31_t)y ) >> 16)) +
- ((((q31_t)x ) >> 16) * (((q31_t)y << 16) >> 16)) +
- ( ((q31_t)sum ) ) ));
- }
-
-
- /*
- * @brief C custom defined SMLSDX
- */
- __STATIC_FORCEINLINE uint32_t __SMLSDX(
- uint32_t x,
- uint32_t y,
- uint32_t sum)
- {
- return ((uint32_t)(((((q31_t)x << 16) >> 16) * (((q31_t)y ) >> 16)) -
- ((((q31_t)x ) >> 16) * (((q31_t)y << 16) >> 16)) +
- ( ((q31_t)sum ) ) ));
- }
-
-
- /*
- * @brief C custom defined SMLALD
- */
- __STATIC_FORCEINLINE uint64_t __SMLALD(
- uint32_t x,
- uint32_t y,
- uint64_t sum)
- {
-/* return (sum + ((q15_t) (x >> 16) * (q15_t) (y >> 16)) + ((q15_t) x * (q15_t) y)); */
- return ((uint64_t)(((((q31_t)x << 16) >> 16) * (((q31_t)y << 16) >> 16)) +
- ((((q31_t)x ) >> 16) * (((q31_t)y ) >> 16)) +
- ( ((q63_t)sum ) ) ));
- }
-
-
- /*
- * @brief C custom defined SMLALDX
- */
- __STATIC_FORCEINLINE uint64_t __SMLALDX(
- uint32_t x,
- uint32_t y,
- uint64_t sum)
- {
-/* return (sum + ((q15_t) (x >> 16) * (q15_t) y)) + ((q15_t) x * (q15_t) (y >> 16)); */
- return ((uint64_t)(((((q31_t)x << 16) >> 16) * (((q31_t)y ) >> 16)) +
- ((((q31_t)x ) >> 16) * (((q31_t)y << 16) >> 16)) +
- ( ((q63_t)sum ) ) ));
- }
-
-
- /*
- * @brief C custom defined SMUAD
- */
- __STATIC_FORCEINLINE uint32_t __SMUAD(
- uint32_t x,
- uint32_t y)
- {
- return ((uint32_t)(((((q31_t)x << 16) >> 16) * (((q31_t)y << 16) >> 16)) +
- ((((q31_t)x ) >> 16) * (((q31_t)y ) >> 16)) ));
- }
-
-
- /*
- * @brief C custom defined SMUSD
- */
- __STATIC_FORCEINLINE uint32_t __SMUSD(
- uint32_t x,
- uint32_t y)
- {
- return ((uint32_t)(((((q31_t)x << 16) >> 16) * (((q31_t)y << 16) >> 16)) -
- ((((q31_t)x ) >> 16) * (((q31_t)y ) >> 16)) ));
- }
-
-
- /*
- * @brief C custom defined SXTB16
- */
- __STATIC_FORCEINLINE uint32_t __SXTB16(
- uint32_t x)
- {
- return ((uint32_t)(((((q31_t)x << 24) >> 24) & (q31_t)0x0000FFFF) |
- ((((q31_t)x << 8) >> 8) & (q31_t)0xFFFF0000) ));
- }
-
- /*
- * @brief C custom defined SMMLA
- */
- __STATIC_FORCEINLINE int32_t __SMMLA(
- int32_t x,
- int32_t y,
- int32_t sum)
- {
- return (sum + (int32_t) (((int64_t) x * y) >> 32));
- }
-
-#endif /* !defined (RISCV_MATH_DSP) */
-
-
- /**
- * @brief Instance structure for the Q7 FIR filter.
- */
- typedef struct
- {
- uint16_t numTaps; /**< number of filter coefficients in the filter. */
- q7_t *pState; /**< points to the state variable array. The array is of length numTaps+blockSize-1. */
- const q7_t *pCoeffs; /**< points to the coefficient array. The array is of length numTaps.*/
- } riscv_fir_instance_q7;
-
- /**
- * @brief Instance structure for the Q15 FIR filter.
- */
- typedef struct
- {
- uint16_t numTaps; /**< number of filter coefficients in the filter. */
- q15_t *pState; /**< points to the state variable array. The array is of length numTaps+blockSize-1. */
- const q15_t *pCoeffs; /**< points to the coefficient array. The array is of length numTaps.*/
- } riscv_fir_instance_q15;
-
- /**
- * @brief Instance structure for the Q31 FIR filter.
- */
- typedef struct
- {
- uint16_t numTaps; /**< number of filter coefficients in the filter. */
- q31_t *pState; /**< points to the state variable array. The array is of length numTaps+blockSize-1. */
- const q31_t *pCoeffs; /**< points to the coefficient array. The array is of length numTaps. */
- } riscv_fir_instance_q31;
-
- /**
- * @brief Instance structure for the floating-point FIR filter.
- */
- typedef struct
- {
- uint16_t numTaps; /**< number of filter coefficients in the filter. */
- float32_t *pState; /**< points to the state variable array. The array is of length numTaps+blockSize-1. */
- const float32_t *pCoeffs; /**< points to the coefficient array. The array is of length numTaps. */
- } riscv_fir_instance_f32;
-
- /**
- * @brief Processing function for the Q7 FIR filter.
- * @param[in] S points to an instance of the Q7 FIR filter structure.
- * @param[in] pSrc points to the block of input data.
- * @param[out] pDst points to the block of output data.
- * @param[in] blockSize number of samples to process.
- */
- void riscv_fir_q7(
- const riscv_fir_instance_q7 * S,
- const q7_t * pSrc,
- q7_t * pDst,
- uint32_t blockSize);
-
- /**
- * @brief Initialization function for the Q7 FIR filter.
- * @param[in,out] S points to an instance of the Q7 FIR structure.
- * @param[in] numTaps Number of filter coefficients in the filter.
- * @param[in] pCoeffs points to the filter coefficients.
- * @param[in] pState points to the state buffer.
- * @param[in] blockSize number of samples that are processed.
- */
- void riscv_fir_init_q7(
- riscv_fir_instance_q7 * S,
- uint16_t numTaps,
- const q7_t * pCoeffs,
- q7_t * pState,
- uint32_t blockSize);
-
- /**
- * @brief Processing function for the Q15 FIR filter.
- * @param[in] S points to an instance of the Q15 FIR structure.
- * @param[in] pSrc points to the block of input data.
- * @param[out] pDst points to the block of output data.
- * @param[in] blockSize number of samples to process.
- */
- void riscv_fir_q15(
- const riscv_fir_instance_q15 * S,
- const q15_t * pSrc,
- q15_t * pDst,
- uint32_t blockSize);
-
- /**
- * @brief Processing function for the fast Q15 FIR filter (fast version).
- * @param[in] S points to an instance of the Q15 FIR filter structure.
- * @param[in] pSrc points to the block of input data.
- * @param[out] pDst points to the block of output data.
- * @param[in] blockSize number of samples to process.
- */
- void riscv_fir_fast_q15(
- const riscv_fir_instance_q15 * S,
- const q15_t * pSrc,
- q15_t * pDst,
- uint32_t blockSize);
-
- /**
- * @brief Initialization function for the Q15 FIR filter.
- * @param[in,out] S points to an instance of the Q15 FIR filter structure.
- * @param[in] numTaps Number of filter coefficients in the filter. Must be even and greater than or equal to 4.
- * @param[in] pCoeffs points to the filter coefficients.
- * @param[in] pState points to the state buffer.
- * @param[in] blockSize number of samples that are processed at a time.
- * @return The function returns either
- * RISCV_MATH_SUCCESS if initialization was successful or
- * RISCV_MATH_ARGUMENT_ERROR if numTaps is not a supported value.
- */
- riscv_status riscv_fir_init_q15(
- riscv_fir_instance_q15 * S,
- uint16_t numTaps,
- const q15_t * pCoeffs,
- q15_t * pState,
- uint32_t blockSize);
-
- /**
- * @brief Processing function for the Q31 FIR filter.
- * @param[in] S points to an instance of the Q31 FIR filter structure.
- * @param[in] pSrc points to the block of input data.
- * @param[out] pDst points to the block of output data.
- * @param[in] blockSize number of samples to process.
- */
- void riscv_fir_q31(
- const riscv_fir_instance_q31 * S,
- const q31_t * pSrc,
- q31_t * pDst,
- uint32_t blockSize);
-
- /**
- * @brief Processing function for the fast Q31 FIR filter (fast version).
- * @param[in] S points to an instance of the Q31 FIR filter structure.
- * @param[in] pSrc points to the block of input data.
- * @param[out] pDst points to the block of output data.
- * @param[in] blockSize number of samples to process.
- */
- void riscv_fir_fast_q31(
- const riscv_fir_instance_q31 * S,
- const q31_t * pSrc,
- q31_t * pDst,
- uint32_t blockSize);
-
- /**
- * @brief Initialization function for the Q31 FIR filter.
- * @param[in,out] S points to an instance of the Q31 FIR structure.
- * @param[in] numTaps Number of filter coefficients in the filter.
- * @param[in] pCoeffs points to the filter coefficients.
- * @param[in] pState points to the state buffer.
- * @param[in] blockSize number of samples that are processed at a time.
- */
- void riscv_fir_init_q31(
- riscv_fir_instance_q31 * S,
- uint16_t numTaps,
- const q31_t * pCoeffs,
- q31_t * pState,
- uint32_t blockSize);
-
- /**
- * @brief Processing function for the floating-point FIR filter.
- * @param[in] S points to an instance of the floating-point FIR structure.
- * @param[in] pSrc points to the block of input data.
- * @param[out] pDst points to the block of output data.
- * @param[in] blockSize number of samples to process.
- */
- void riscv_fir_f32(
- const riscv_fir_instance_f32 * S,
- const float32_t * pSrc,
- float32_t * pDst,
- uint32_t blockSize);
-
- /**
- * @brief Initialization function for the floating-point FIR filter.
- * @param[in,out] S points to an instance of the floating-point FIR filter structure.
- * @param[in] numTaps Number of filter coefficients in the filter.
- * @param[in] pCoeffs points to the filter coefficients.
- * @param[in] pState points to the state buffer.
- * @param[in] blockSize number of samples that are processed at a time.
- */
- void riscv_fir_init_f32(
- riscv_fir_instance_f32 * S,
- uint16_t numTaps,
- const float32_t * pCoeffs,
- float32_t * pState,
- uint32_t blockSize);
-
- /**
- * @brief Instance structure for the Q15 Biquad cascade filter.
- */
- typedef struct
- {
- int8_t numStages; /**< number of 2nd order stages in the filter. Overall order is 2*numStages. */
- q15_t *pState; /**< Points to the array of state coefficients. The array is of length 4*numStages. */
- const q15_t *pCoeffs; /**< Points to the array of coefficients. The array is of length 5*numStages. */
- int8_t postShift; /**< Additional shift, in bits, applied to each output sample. */
- } riscv_biquad_casd_df1_inst_q15;
-
- /**
- * @brief Instance structure for the Q31 Biquad cascade filter.
- */
- typedef struct
- {
- uint32_t numStages; /**< number of 2nd order stages in the filter. Overall order is 2*numStages. */
- q31_t *pState; /**< Points to the array of state coefficients. The array is of length 4*numStages. */
- const q31_t *pCoeffs; /**< Points to the array of coefficients. The array is of length 5*numStages. */
- uint8_t postShift; /**< Additional shift, in bits, applied to each output sample. */
- } riscv_biquad_casd_df1_inst_q31;
-
- /**
- * @brief Instance structure for the floating-point Biquad cascade filter.
- */
- typedef struct
- {
- uint32_t numStages; /**< number of 2nd order stages in the filter. Overall order is 2*numStages. */
- float32_t *pState; /**< Points to the array of state coefficients. The array is of length 4*numStages. */
- const float32_t *pCoeffs; /**< Points to the array of coefficients. The array is of length 5*numStages. */
- } riscv_biquad_casd_df1_inst_f32;
-
- /**
- * @brief Processing function for the Q15 Biquad cascade filter.
- * @param[in] S points to an instance of the Q15 Biquad cascade structure.
- * @param[in] pSrc points to the block of input data.
- * @param[out] pDst points to the block of output data.
- * @param[in] blockSize number of samples to process.
- */
- void riscv_biquad_cascade_df1_q15(
- const riscv_biquad_casd_df1_inst_q15 * S,
- const q15_t * pSrc,
- q15_t * pDst,
- uint32_t blockSize);
-
- /**
- * @brief Initialization function for the Q15 Biquad cascade filter.
- * @param[in,out] S points to an instance of the Q15 Biquad cascade structure.
- * @param[in] numStages number of 2nd order stages in the filter.
- * @param[in] pCoeffs points to the filter coefficients.
- * @param[in] pState points to the state buffer.
- * @param[in] postShift Shift to be applied to the output. Varies according to the coefficients format
- */
- void riscv_biquad_cascade_df1_init_q15(
- riscv_biquad_casd_df1_inst_q15 * S,
- uint8_t numStages,
- const q15_t * pCoeffs,
- q15_t * pState,
- int8_t postShift);
-
- /**
- * @brief Fast but less precise processing function for the Q15 Biquad cascade filter for RISC-V Core with DSP enabled.
- * @param[in] S points to an instance of the Q15 Biquad cascade structure.
- * @param[in] pSrc points to the block of input data.
- * @param[out] pDst points to the block of output data.
- * @param[in] blockSize number of samples to process.
- */
- void riscv_biquad_cascade_df1_fast_q15(
- const riscv_biquad_casd_df1_inst_q15 * S,
- const q15_t * pSrc,
- q15_t * pDst,
- uint32_t blockSize);
-
- /**
- * @brief Processing function for the Q31 Biquad cascade filter
- * @param[in] S points to an instance of the Q31 Biquad cascade structure.
- * @param[in] pSrc points to the block of input data.
- * @param[out] pDst points to the block of output data.
- * @param[in] blockSize number of samples to process.
- */
- void riscv_biquad_cascade_df1_q31(
- const riscv_biquad_casd_df1_inst_q31 * S,
- const q31_t * pSrc,
- q31_t * pDst,
- uint32_t blockSize);
-
- /**
- * @brief Fast but less precise processing function for the Q31 Biquad cascade filter for RISC-V Core with DSP enabled.
- * @param[in] S points to an instance of the Q31 Biquad cascade structure.
- * @param[in] pSrc points to the block of input data.
- * @param[out] pDst points to the block of output data.
- * @param[in] blockSize number of samples to process.
- */
- void riscv_biquad_cascade_df1_fast_q31(
- const riscv_biquad_casd_df1_inst_q31 * S,
- const q31_t * pSrc,
- q31_t * pDst,
- uint32_t blockSize);
-
- /**
- * @brief Initialization function for the Q31 Biquad cascade filter.
- * @param[in,out] S points to an instance of the Q31 Biquad cascade structure.
- * @param[in] numStages number of 2nd order stages in the filter.
- * @param[in] pCoeffs points to the filter coefficients.
- * @param[in] pState points to the state buffer.
- * @param[in] postShift Shift to be applied to the output. Varies according to the coefficients format
- */
- void riscv_biquad_cascade_df1_init_q31(
- riscv_biquad_casd_df1_inst_q31 * S,
- uint8_t numStages,
- const q31_t * pCoeffs,
- q31_t * pState,
- int8_t postShift);
-
- /**
- * @brief Processing function for the floating-point Biquad cascade filter.
- * @param[in] S points to an instance of the floating-point Biquad cascade structure.
- * @param[in] pSrc points to the block of input data.
- * @param[out] pDst points to the block of output data.
- * @param[in] blockSize number of samples to process.
- */
- void riscv_biquad_cascade_df1_f32(
- const riscv_biquad_casd_df1_inst_f32 * S,
- const float32_t * pSrc,
- float32_t * pDst,
- uint32_t blockSize);
-
- /**
- * @brief Initialization function for the floating-point Biquad cascade filter.
- * @param[in,out] S points to an instance of the floating-point Biquad cascade structure.
- * @param[in] numStages number of 2nd order stages in the filter.
- * @param[in] pCoeffs points to the filter coefficients.
- * @param[in] pState points to the state buffer.
- */
- void riscv_biquad_cascade_df1_init_f32(
- riscv_biquad_casd_df1_inst_f32 * S,
- uint8_t numStages,
- const float32_t * pCoeffs,
- float32_t * pState);
-
- /**
- * @brief Instance structure for the floating-point matrix structure.
- */
- typedef struct
- {
- uint16_t numRows; /**< number of rows of the matrix. */
- uint16_t numCols; /**< number of columns of the matrix. */
- float32_t *pData; /**< points to the data of the matrix. */
- } riscv_matrix_instance_f32;
-
-
- /**
- * @brief Instance structure for the floating-point matrix structure.
- */
- typedef struct
- {
- uint16_t numRows; /**< number of rows of the matrix. */
- uint16_t numCols; /**< number of columns of the matrix. */
- float64_t *pData; /**< points to the data of the matrix. */
- } riscv_matrix_instance_f64;
-
- /**
- * @brief Instance structure for the Q15 matrix structure.
- */
- typedef struct
- {
- uint16_t numRows; /**< number of rows of the matrix. */
- uint16_t numCols; /**< number of columns of the matrix. */
- q15_t *pData; /**< points to the data of the matrix. */
- } riscv_matrix_instance_q15;
-
- /**
- * @brief Instance structure for the Q31 matrix structure.
- */
- typedef struct
- {
- uint16_t numRows; /**< number of rows of the matrix. */
- uint16_t numCols; /**< number of columns of the matrix. */
- q31_t *pData; /**< points to the data of the matrix. */
- } riscv_matrix_instance_q31;
-
- /**
- * @brief Floating-point matrix addition.
- * @param[in] pSrcA points to the first input matrix structure
- * @param[in] pSrcB points to the second input matrix structure
- * @param[out] pDst points to output matrix structure
- * @return The function returns either
- * RISCV_MATH_SIZE_MISMATCH or RISCV_MATH_SUCCESS based on the outcome of size checking.
- */
-riscv_status riscv_mat_add_f32(
- const riscv_matrix_instance_f32 * pSrcA,
- const riscv_matrix_instance_f32 * pSrcB,
- riscv_matrix_instance_f32 * pDst);
-
- /**
- * @brief Q15 matrix addition.
- * @param[in] pSrcA points to the first input matrix structure
- * @param[in] pSrcB points to the second input matrix structure
- * @param[out] pDst points to output matrix structure
- * @return The function returns either
- * RISCV_MATH_SIZE_MISMATCH or RISCV_MATH_SUCCESS based on the outcome of size checking.
- */
-riscv_status riscv_mat_add_q15(
- const riscv_matrix_instance_q15 * pSrcA,
- const riscv_matrix_instance_q15 * pSrcB,
- riscv_matrix_instance_q15 * pDst);
-
- /**
- * @brief Q31 matrix addition.
- * @param[in] pSrcA points to the first input matrix structure
- * @param[in] pSrcB points to the second input matrix structure
- * @param[out] pDst points to output matrix structure
- * @return The function returns either
- * RISCV_MATH_SIZE_MISMATCH or RISCV_MATH_SUCCESS based on the outcome of size checking.
- */
-riscv_status riscv_mat_add_q31(
- const riscv_matrix_instance_q31 * pSrcA,
- const riscv_matrix_instance_q31 * pSrcB,
- riscv_matrix_instance_q31 * pDst);
-
- /**
- * @brief Floating-point, complex, matrix multiplication.
- * @param[in] pSrcA points to the first input matrix structure
- * @param[in] pSrcB points to the second input matrix structure
- * @param[out] pDst points to output matrix structure
- * @return The function returns either
- * RISCV_MATH_SIZE_MISMATCH or RISCV_MATH_SUCCESS based on the outcome of size checking.
- */
-riscv_status riscv_mat_cmplx_mult_f32(
- const riscv_matrix_instance_f32 * pSrcA,
- const riscv_matrix_instance_f32 * pSrcB,
- riscv_matrix_instance_f32 * pDst);
-
- /**
- * @brief Q15, complex, matrix multiplication.
- * @param[in] pSrcA points to the first input matrix structure
- * @param[in] pSrcB points to the second input matrix structure
- * @param[out] pDst points to output matrix structure
- * @return The function returns either
- * RISCV_MATH_SIZE_MISMATCH or RISCV_MATH_SUCCESS based on the outcome of size checking.
- */
-riscv_status riscv_mat_cmplx_mult_q15(
- const riscv_matrix_instance_q15 * pSrcA,
- const riscv_matrix_instance_q15 * pSrcB,
- riscv_matrix_instance_q15 * pDst,
- q15_t * pScratch);
-
- /**
- * @brief Q31, complex, matrix multiplication.
- * @param[in] pSrcA points to the first input matrix structure
- * @param[in] pSrcB points to the second input matrix structure
- * @param[out] pDst points to output matrix structure
- * @return The function returns either
- * RISCV_MATH_SIZE_MISMATCH or RISCV_MATH_SUCCESS based on the outcome of size checking.
- */
-riscv_status riscv_mat_cmplx_mult_q31(
- const riscv_matrix_instance_q31 * pSrcA,
- const riscv_matrix_instance_q31 * pSrcB,
- riscv_matrix_instance_q31 * pDst);
-
- /**
- * @brief Floating-point matrix transpose.
- * @param[in] pSrc points to the input matrix
- * @param[out] pDst points to the output matrix
- * @return The function returns either RISCV_MATH_SIZE_MISMATCH
- * or RISCV_MATH_SUCCESS based on the outcome of size checking.
- */
-riscv_status riscv_mat_trans_f32(
- const riscv_matrix_instance_f32 * pSrc,
- riscv_matrix_instance_f32 * pDst);
-
- /**
- * @brief Q15 matrix transpose.
- * @param[in] pSrc points to the input matrix
- * @param[out] pDst points to the output matrix
- * @return The function returns either RISCV_MATH_SIZE_MISMATCH
- * or RISCV_MATH_SUCCESS based on the outcome of size checking.
- */
-riscv_status riscv_mat_trans_q15(
- const riscv_matrix_instance_q15 * pSrc,
- riscv_matrix_instance_q15 * pDst);
-
- /**
- * @brief Q31 matrix transpose.
- * @param[in] pSrc points to the input matrix
- * @param[out] pDst points to the output matrix
- * @return The function returns either RISCV_MATH_SIZE_MISMATCH
- * or RISCV_MATH_SUCCESS based on the outcome of size checking.
- */
-riscv_status riscv_mat_trans_q31(
- const riscv_matrix_instance_q31 * pSrc,
- riscv_matrix_instance_q31 * pDst);
-
- /**
- * @brief Floating-point matrix multiplication
- * @param[in] pSrcA points to the first input matrix structure
- * @param[in] pSrcB points to the second input matrix structure
- * @param[out] pDst points to output matrix structure
- * @return The function returns either
- * RISCV_MATH_SIZE_MISMATCH or RISCV_MATH_SUCCESS based on the outcome of size checking.
- */
-riscv_status riscv_mat_mult_f32(
- const riscv_matrix_instance_f32 * pSrcA,
- const riscv_matrix_instance_f32 * pSrcB,
- riscv_matrix_instance_f32 * pDst);
-
- /**
- * @brief Q15 matrix multiplication
- * @param[in] pSrcA points to the first input matrix structure
- * @param[in] pSrcB points to the second input matrix structure
- * @param[out] pDst points to output matrix structure
- * @param[in] pState points to the array for storing intermediate results
- * @return The function returns either
- * RISCV_MATH_SIZE_MISMATCH or RISCV_MATH_SUCCESS based on the outcome of size checking.
- */
-riscv_status riscv_mat_mult_q15(
- const riscv_matrix_instance_q15 * pSrcA,
- const riscv_matrix_instance_q15 * pSrcB,
- riscv_matrix_instance_q15 * pDst,
- q15_t * pState);
-
- /**
- * @brief Q15 matrix multiplication (fast variant) for RISC-V Core with DSP enabled
- * @param[in] pSrcA points to the first input matrix structure
- * @param[in] pSrcB points to the second input matrix structure
- * @param[out] pDst points to output matrix structure
- * @param[in] pState points to the array for storing intermediate results
- * @return The function returns either
- * RISCV_MATH_SIZE_MISMATCH or RISCV_MATH_SUCCESS based on the outcome of size checking.
- */
-riscv_status riscv_mat_mult_fast_q15(
- const riscv_matrix_instance_q15 * pSrcA,
- const riscv_matrix_instance_q15 * pSrcB,
- riscv_matrix_instance_q15 * pDst,
- q15_t * pState);
-
- /**
- * @brief Q31 matrix multiplication
- * @param[in] pSrcA points to the first input matrix structure
- * @param[in] pSrcB points to the second input matrix structure
- * @param[out] pDst points to output matrix structure
- * @return The function returns either
- * RISCV_MATH_SIZE_MISMATCH or RISCV_MATH_SUCCESS based on the outcome of size checking.
- */
-riscv_status riscv_mat_mult_q31(
- const riscv_matrix_instance_q31 * pSrcA,
- const riscv_matrix_instance_q31 * pSrcB,
- riscv_matrix_instance_q31 * pDst);
-
- /**
- * @brief Q31 matrix multiplication (fast variant) for RISC-V Core with DSP enabled
- * @param[in] pSrcA points to the first input matrix structure
- * @param[in] pSrcB points to the second input matrix structure
- * @param[out] pDst points to output matrix structure
- * @return The function returns either
- * RISCV_MATH_SIZE_MISMATCH or RISCV_MATH_SUCCESS based on the outcome of size checking.
- */
-riscv_status riscv_mat_mult_fast_q31(
- const riscv_matrix_instance_q31 * pSrcA,
- const riscv_matrix_instance_q31 * pSrcB,
- riscv_matrix_instance_q31 * pDst);
-
- /**
- * @brief Floating-point matrix subtraction
- * @param[in] pSrcA points to the first input matrix structure
- * @param[in] pSrcB points to the second input matrix structure
- * @param[out] pDst points to output matrix structure
- * @return The function returns either
- * RISCV_MATH_SIZE_MISMATCH or RISCV_MATH_SUCCESS based on the outcome of size checking.
- */
-riscv_status riscv_mat_sub_f32(
- const riscv_matrix_instance_f32 * pSrcA,
- const riscv_matrix_instance_f32 * pSrcB,
- riscv_matrix_instance_f32 * pDst);
-
- /**
- * @brief Q15 matrix subtraction
- * @param[in] pSrcA points to the first input matrix structure
- * @param[in] pSrcB points to the second input matrix structure
- * @param[out] pDst points to output matrix structure
- * @return The function returns either
- * RISCV_MATH_SIZE_MISMATCH or RISCV_MATH_SUCCESS based on the outcome of size checking.
- */
-riscv_status riscv_mat_sub_q15(
- const riscv_matrix_instance_q15 * pSrcA,
- const riscv_matrix_instance_q15 * pSrcB,
- riscv_matrix_instance_q15 * pDst);
-
- /**
- * @brief Q31 matrix subtraction
- * @param[in] pSrcA points to the first input matrix structure
- * @param[in] pSrcB points to the second input matrix structure
- * @param[out] pDst points to output matrix structure
- * @return The function returns either
- * RISCV_MATH_SIZE_MISMATCH or RISCV_MATH_SUCCESS based on the outcome of size checking.
- */
-riscv_status riscv_mat_sub_q31(
- const riscv_matrix_instance_q31 * pSrcA,
- const riscv_matrix_instance_q31 * pSrcB,
- riscv_matrix_instance_q31 * pDst);
-
- /**
- * @brief Floating-point matrix scaling.
- * @param[in] pSrc points to the input matrix
- * @param[in] scale scale factor
- * @param[out] pDst points to the output matrix
- * @return The function returns either
- * RISCV_MATH_SIZE_MISMATCH or RISCV_MATH_SUCCESS based on the outcome of size checking.
- */
-riscv_status riscv_mat_scale_f32(
- const riscv_matrix_instance_f32 * pSrc,
- float32_t scale,
- riscv_matrix_instance_f32 * pDst);
-
- /**
- * @brief Q15 matrix scaling.
- * @param[in] pSrc points to input matrix
- * @param[in] scaleFract fractional portion of the scale factor
- * @param[in] shift number of bits to shift the result by
- * @param[out] pDst points to output matrix
- * @return The function returns either
- * RISCV_MATH_SIZE_MISMATCH or RISCV_MATH_SUCCESS based on the outcome of size checking.
- */
-riscv_status riscv_mat_scale_q15(
- const riscv_matrix_instance_q15 * pSrc,
- q15_t scaleFract,
- int32_t shift,
- riscv_matrix_instance_q15 * pDst);
-
- /**
- * @brief Q31 matrix scaling.
- * @param[in] pSrc points to input matrix
- * @param[in] scaleFract fractional portion of the scale factor
- * @param[in] shift number of bits to shift the result by
- * @param[out] pDst points to output matrix structure
- * @return The function returns either
- * RISCV_MATH_SIZE_MISMATCH or RISCV_MATH_SUCCESS based on the outcome of size checking.
- */
-riscv_status riscv_mat_scale_q31(
- const riscv_matrix_instance_q31 * pSrc,
- q31_t scaleFract,
- int32_t shift,
- riscv_matrix_instance_q31 * pDst);
-
- /**
- * @brief Q31 matrix initialization.
- * @param[in,out] S points to an instance of the floating-point matrix structure.
- * @param[in] nRows number of rows in the matrix.
- * @param[in] nColumns number of columns in the matrix.
- * @param[in] pData points to the matrix data array.
- */
-void riscv_mat_init_q31(
- riscv_matrix_instance_q31 * S,
- uint16_t nRows,
- uint16_t nColumns,
- q31_t * pData);
-
- /**
- * @brief Q15 matrix initialization.
- * @param[in,out] S points to an instance of the floating-point matrix structure.
- * @param[in] nRows number of rows in the matrix.
- * @param[in] nColumns number of columns in the matrix.
- * @param[in] pData points to the matrix data array.
- */
-void riscv_mat_init_q15(
- riscv_matrix_instance_q15 * S,
- uint16_t nRows,
- uint16_t nColumns,
- q15_t * pData);
-
- /**
- * @brief Floating-point matrix initialization.
- * @param[in,out] S points to an instance of the floating-point matrix structure.
- * @param[in] nRows number of rows in the matrix.
- * @param[in] nColumns number of columns in the matrix.
- * @param[in] pData points to the matrix data array.
- */
-void riscv_mat_init_f32(
- riscv_matrix_instance_f32 * S,
- uint16_t nRows,
- uint16_t nColumns,
- float32_t * pData);
-
-
- /**
- * @brief Instance structure for the Q15 PID Control.
- */
- typedef struct
- {
- q15_t A0; /**< The derived gain, A0 = Kp + Ki + Kd . */
-#if !defined (RISCV_MATH_DSP)
- q15_t A1;
- q15_t A2;
-#else
- q31_t A1; /**< The derived gain A1 = -Kp - 2Kd | Kd.*/
-#endif
- q15_t state[3]; /**< The state array of length 3. */
- q15_t Kp; /**< The proportional gain. */
- q15_t Ki; /**< The integral gain. */
- q15_t Kd; /**< The derivative gain. */
- } riscv_pid_instance_q15;
-
- /**
- * @brief Instance structure for the Q31 PID Control.
- */
- typedef struct
- {
- q31_t A0; /**< The derived gain, A0 = Kp + Ki + Kd . */
- q31_t A1; /**< The derived gain, A1 = -Kp - 2Kd. */
- q31_t A2; /**< The derived gain, A2 = Kd . */
- q31_t state[3]; /**< The state array of length 3. */
- q31_t Kp; /**< The proportional gain. */
- q31_t Ki; /**< The integral gain. */
- q31_t Kd; /**< The derivative gain. */
- } riscv_pid_instance_q31;
-
- /**
- * @brief Instance structure for the floating-point PID Control.
- */
- typedef struct
- {
- float32_t A0; /**< The derived gain, A0 = Kp + Ki + Kd . */
- float32_t A1; /**< The derived gain, A1 = -Kp - 2Kd. */
- float32_t A2; /**< The derived gain, A2 = Kd . */
- float32_t state[3]; /**< The state array of length 3. */
- float32_t Kp; /**< The proportional gain. */
- float32_t Ki; /**< The integral gain. */
- float32_t Kd; /**< The derivative gain. */
- } riscv_pid_instance_f32;
-
-
-
- /**
- * @brief Initialization function for the floating-point PID Control.
- * @param[in,out] S points to an instance of the PID structure.
- * @param[in] resetStateFlag flag to reset the state. 0 = no change in state 1 = reset the state.
- */
- void riscv_pid_init_f32(
- riscv_pid_instance_f32 * S,
- int32_t resetStateFlag);
-
-
- /**
- * @brief Reset function for the floating-point PID Control.
- * @param[in,out] S is an instance of the floating-point PID Control structure
- */
- void riscv_pid_reset_f32(
- riscv_pid_instance_f32 * S);
-
-
- /**
- * @brief Initialization function for the Q31 PID Control.
- * @param[in,out] S points to an instance of the Q15 PID structure.
- * @param[in] resetStateFlag flag to reset the state. 0 = no change in state 1 = reset the state.
- */
- void riscv_pid_init_q31(
- riscv_pid_instance_q31 * S,
- int32_t resetStateFlag);
-
-
- /**
- * @brief Reset function for the Q31 PID Control.
- * @param[in,out] S points to an instance of the Q31 PID Control structure
- */
-
- void riscv_pid_reset_q31(
- riscv_pid_instance_q31 * S);
-
-
- /**
- * @brief Initialization function for the Q15 PID Control.
- * @param[in,out] S points to an instance of the Q15 PID structure.
- * @param[in] resetStateFlag flag to reset the state. 0 = no change in state 1 = reset the state.
- */
- void riscv_pid_init_q15(
- riscv_pid_instance_q15 * S,
- int32_t resetStateFlag);
-
-
- /**
- * @brief Reset function for the Q15 PID Control.
- * @param[in,out] S points to an instance of the q15 PID Control structure
- */
- void riscv_pid_reset_q15(
- riscv_pid_instance_q15 * S);
-
-
- /**
- * @brief Instance structure for the floating-point Linear Interpolate function.
- */
- typedef struct
- {
- uint32_t nValues; /**< nValues */
- float32_t x1; /**< x1 */
- float32_t xSpacing; /**< xSpacing */
- float32_t *pYData; /**< pointer to the table of Y values */
- } riscv_linear_interp_instance_f32;
-
- /**
- * @brief Instance structure for the floating-point bilinear interpolation function.
- */
- typedef struct
- {
- uint16_t numRows; /**< number of rows in the data table. */
- uint16_t numCols; /**< number of columns in the data table. */
- float32_t *pData; /**< points to the data table. */
- } riscv_bilinear_interp_instance_f32;
-
- /**
- * @brief Instance structure for the Q31 bilinear interpolation function.
- */
- typedef struct
- {
- uint16_t numRows; /**< number of rows in the data table. */
- uint16_t numCols; /**< number of columns in the data table. */
- q31_t *pData; /**< points to the data table. */
- } riscv_bilinear_interp_instance_q31;
-
- /**
- * @brief Instance structure for the Q15 bilinear interpolation function.
- */
- typedef struct
- {
- uint16_t numRows; /**< number of rows in the data table. */
- uint16_t numCols; /**< number of columns in the data table. */
- q15_t *pData; /**< points to the data table. */
- } riscv_bilinear_interp_instance_q15;
-
- /**
- * @brief Instance structure for the Q15 bilinear interpolation function.
- */
- typedef struct
- {
- uint16_t numRows; /**< number of rows in the data table. */
- uint16_t numCols; /**< number of columns in the data table. */
- q7_t *pData; /**< points to the data table. */
- } riscv_bilinear_interp_instance_q7;
-
-
- /**
- * @brief Q7 vector multiplication.
- * @param[in] pSrcA points to the first input vector
- * @param[in] pSrcB points to the second input vector
- * @param[out] pDst points to the output vector
- * @param[in] blockSize number of samples in each vector
- */
- void riscv_mult_q7(
- const q7_t * pSrcA,
- const q7_t * pSrcB,
- q7_t * pDst,
- uint32_t blockSize);
-
-
- /**
- * @brief Q15 vector multiplication.
- * @param[in] pSrcA points to the first input vector
- * @param[in] pSrcB points to the second input vector
- * @param[out] pDst points to the output vector
- * @param[in] blockSize number of samples in each vector
- */
- void riscv_mult_q15(
- const q15_t * pSrcA,
- const q15_t * pSrcB,
- q15_t * pDst,
- uint32_t blockSize);
-
-
- /**
- * @brief Q31 vector multiplication.
- * @param[in] pSrcA points to the first input vector
- * @param[in] pSrcB points to the second input vector
- * @param[out] pDst points to the output vector
- * @param[in] blockSize number of samples in each vector
- */
- void riscv_mult_q31(
- const q31_t * pSrcA,
- const q31_t * pSrcB,
- q31_t * pDst,
- uint32_t blockSize);
-
-
- /**
- * @brief Floating-point vector multiplication.
- * @param[in] pSrcA points to the first input vector
- * @param[in] pSrcB points to the second input vector
- * @param[out] pDst points to the output vector
- * @param[in] blockSize number of samples in each vector
- */
- void riscv_mult_f32(
- const float32_t * pSrcA,
- const float32_t * pSrcB,
- float32_t * pDst,
- uint32_t blockSize);
-
-
- /**
- * @brief Instance structure for the Q15 CFFT/CIFFT function.
- */
- typedef struct
- {
- uint16_t fftLen; /**< length of the FFT. */
- uint8_t ifftFlag; /**< flag that selects forward (ifftFlag=0) or inverse (ifftFlag=1) transform. */
- uint8_t bitReverseFlag; /**< flag that enables (bitReverseFlag=1) or disables (bitReverseFlag=0) bit reversal of output. */
- const q15_t *pTwiddle; /**< points to the Sin twiddle factor table. */
- const uint16_t *pBitRevTable; /**< points to the bit reversal table. */
- uint16_t twidCoefModifier; /**< twiddle coefficient modifier that supports different size FFTs with the same twiddle factor table. */
- uint16_t bitRevFactor; /**< bit reversal modifier that supports different size FFTs with the same bit reversal table. */
- } riscv_cfft_radix2_instance_q15;
-
-/* Deprecated */
- riscv_status riscv_cfft_radix2_init_q15(
- riscv_cfft_radix2_instance_q15 * S,
- uint16_t fftLen,
- uint8_t ifftFlag,
- uint8_t bitReverseFlag);
-
-/* Deprecated */
- void riscv_cfft_radix2_q15(
- const riscv_cfft_radix2_instance_q15 * S,
- q15_t * pSrc);
-
-
- /**
- * @brief Instance structure for the Q15 CFFT/CIFFT function.
- */
- typedef struct
- {
- uint16_t fftLen; /**< length of the FFT. */
- uint8_t ifftFlag; /**< flag that selects forward (ifftFlag=0) or inverse (ifftFlag=1) transform. */
- uint8_t bitReverseFlag; /**< flag that enables (bitReverseFlag=1) or disables (bitReverseFlag=0) bit reversal of output. */
- const q15_t *pTwiddle; /**< points to the twiddle factor table. */
- const uint16_t *pBitRevTable; /**< points to the bit reversal table. */
- uint16_t twidCoefModifier; /**< twiddle coefficient modifier that supports different size FFTs with the same twiddle factor table. */
- uint16_t bitRevFactor; /**< bit reversal modifier that supports different size FFTs with the same bit reversal table. */
- } riscv_cfft_radix4_instance_q15;
-
-/* Deprecated */
- riscv_status riscv_cfft_radix4_init_q15(
- riscv_cfft_radix4_instance_q15 * S,
- uint16_t fftLen,
- uint8_t ifftFlag,
- uint8_t bitReverseFlag);
-
-/* Deprecated */
- void riscv_cfft_radix4_q15(
- const riscv_cfft_radix4_instance_q15 * S,
- q15_t * pSrc);
-
- /**
- * @brief Instance structure for the Radix-2 Q31 CFFT/CIFFT function.
- */
- typedef struct
- {
- uint16_t fftLen; /**< length of the FFT. */
- uint8_t ifftFlag; /**< flag that selects forward (ifftFlag=0) or inverse (ifftFlag=1) transform. */
- uint8_t bitReverseFlag; /**< flag that enables (bitReverseFlag=1) or disables (bitReverseFlag=0) bit reversal of output. */
- const q31_t *pTwiddle; /**< points to the Twiddle factor table. */
- const uint16_t *pBitRevTable; /**< points to the bit reversal table. */
- uint16_t twidCoefModifier; /**< twiddle coefficient modifier that supports different size FFTs with the same twiddle factor table. */
- uint16_t bitRevFactor; /**< bit reversal modifier that supports different size FFTs with the same bit reversal table. */
- } riscv_cfft_radix2_instance_q31;
-
-/* Deprecated */
- riscv_status riscv_cfft_radix2_init_q31(
- riscv_cfft_radix2_instance_q31 * S,
- uint16_t fftLen,
- uint8_t ifftFlag,
- uint8_t bitReverseFlag);
-
-/* Deprecated */
- void riscv_cfft_radix2_q31(
- const riscv_cfft_radix2_instance_q31 * S,
- q31_t * pSrc);
-
- /**
- * @brief Instance structure for the Q31 CFFT/CIFFT function.
- */
- typedef struct
- {
- uint16_t fftLen; /**< length of the FFT. */
- uint8_t ifftFlag; /**< flag that selects forward (ifftFlag=0) or inverse (ifftFlag=1) transform. */
- uint8_t bitReverseFlag; /**< flag that enables (bitReverseFlag=1) or disables (bitReverseFlag=0) bit reversal of output. */
- const q31_t *pTwiddle; /**< points to the twiddle factor table. */
- const uint16_t *pBitRevTable; /**< points to the bit reversal table. */
- uint16_t twidCoefModifier; /**< twiddle coefficient modifier that supports different size FFTs with the same twiddle factor table. */
- uint16_t bitRevFactor; /**< bit reversal modifier that supports different size FFTs with the same bit reversal table. */
- } riscv_cfft_radix4_instance_q31;
-
-/* Deprecated */
- void riscv_cfft_radix4_q31(
- const riscv_cfft_radix4_instance_q31 * S,
- q31_t * pSrc);
-
-/* Deprecated */
- riscv_status riscv_cfft_radix4_init_q31(
- riscv_cfft_radix4_instance_q31 * S,
- uint16_t fftLen,
- uint8_t ifftFlag,
- uint8_t bitReverseFlag);
-
- /**
- * @brief Instance structure for the floating-point CFFT/CIFFT function.
- */
- typedef struct
- {
- uint16_t fftLen; /**< length of the FFT. */
- uint8_t ifftFlag; /**< flag that selects forward (ifftFlag=0) or inverse (ifftFlag=1) transform. */
- uint8_t bitReverseFlag; /**< flag that enables (bitReverseFlag=1) or disables (bitReverseFlag=0) bit reversal of output. */
- const float32_t *pTwiddle; /**< points to the Twiddle factor table. */
- const uint16_t *pBitRevTable; /**< points to the bit reversal table. */
- uint16_t twidCoefModifier; /**< twiddle coefficient modifier that supports different size FFTs with the same twiddle factor table. */
- uint16_t bitRevFactor; /**< bit reversal modifier that supports different size FFTs with the same bit reversal table. */
- float32_t onebyfftLen; /**< value of 1/fftLen. */
- } riscv_cfft_radix2_instance_f32;
-
-/* Deprecated */
- riscv_status riscv_cfft_radix2_init_f32(
- riscv_cfft_radix2_instance_f32 * S,
- uint16_t fftLen,
- uint8_t ifftFlag,
- uint8_t bitReverseFlag);
-
-/* Deprecated */
- void riscv_cfft_radix2_f32(
- const riscv_cfft_radix2_instance_f32 * S,
- float32_t * pSrc);
-
- /**
- * @brief Instance structure for the floating-point CFFT/CIFFT function.
- */
- typedef struct
- {
- uint16_t fftLen; /**< length of the FFT. */
- uint8_t ifftFlag; /**< flag that selects forward (ifftFlag=0) or inverse (ifftFlag=1) transform. */
- uint8_t bitReverseFlag; /**< flag that enables (bitReverseFlag=1) or disables (bitReverseFlag=0) bit reversal of output. */
- const float32_t *pTwiddle; /**< points to the Twiddle factor table. */
- const uint16_t *pBitRevTable; /**< points to the bit reversal table. */
- uint16_t twidCoefModifier; /**< twiddle coefficient modifier that supports different size FFTs with the same twiddle factor table. */
- uint16_t bitRevFactor; /**< bit reversal modifier that supports different size FFTs with the same bit reversal table. */
- float32_t onebyfftLen; /**< value of 1/fftLen. */
- } riscv_cfft_radix4_instance_f32;
-
-/* Deprecated */
- riscv_status riscv_cfft_radix4_init_f32(
- riscv_cfft_radix4_instance_f32 * S,
- uint16_t fftLen,
- uint8_t ifftFlag,
- uint8_t bitReverseFlag);
-
-/* Deprecated */
- void riscv_cfft_radix4_f32(
- const riscv_cfft_radix4_instance_f32 * S,
- float32_t * pSrc);
-
- /**
- * @brief Instance structure for the fixed-point CFFT/CIFFT function.
- */
- typedef struct
- {
- uint16_t fftLen; /**< length of the FFT. */
- const q15_t *pTwiddle; /**< points to the Twiddle factor table. */
- const uint16_t *pBitRevTable; /**< points to the bit reversal table. */
- uint16_t bitRevLength; /**< bit reversal table length. */
- } riscv_cfft_instance_q15;
-
-void riscv_cfft_q15(
- const riscv_cfft_instance_q15 * S,
- q15_t * p1,
- uint8_t ifftFlag,
- uint8_t bitReverseFlag);
-
- /**
- * @brief Instance structure for the fixed-point CFFT/CIFFT function.
- */
- typedef struct
- {
- uint16_t fftLen; /**< length of the FFT. */
- const q31_t *pTwiddle; /**< points to the Twiddle factor table. */
- const uint16_t *pBitRevTable; /**< points to the bit reversal table. */
- uint16_t bitRevLength; /**< bit reversal table length. */
- } riscv_cfft_instance_q31;
-
-void riscv_cfft_q31(
- const riscv_cfft_instance_q31 * S,
- q31_t * p1,
- uint8_t ifftFlag,
- uint8_t bitReverseFlag);
-
- /**
- * @brief Instance structure for the floating-point CFFT/CIFFT function.
- */
- typedef struct
- {
- uint16_t fftLen; /**< length of the FFT. */
- const float32_t *pTwiddle; /**< points to the Twiddle factor table. */
- const uint16_t *pBitRevTable; /**< points to the bit reversal table. */
- uint16_t bitRevLength; /**< bit reversal table length. */
- } riscv_cfft_instance_f32;
-
- void riscv_cfft_f32(
- const riscv_cfft_instance_f32 * S,
- float32_t * p1,
- uint8_t ifftFlag,
- uint8_t bitReverseFlag);
-
- /**
- * @brief Instance structure for the Q15 RFFT/RIFFT function.
- */
- typedef struct
- {
- uint32_t fftLenReal; /**< length of the real FFT. */
- uint8_t ifftFlagR; /**< flag that selects forward (ifftFlagR=0) or inverse (ifftFlagR=1) transform. */
- uint8_t bitReverseFlagR; /**< flag that enables (bitReverseFlagR=1) or disables (bitReverseFlagR=0) bit reversal of output. */
- uint32_t twidCoefRModifier; /**< twiddle coefficient modifier that supports different size FFTs with the same twiddle factor table. */
- const q15_t *pTwiddleAReal; /**< points to the real twiddle factor table. */
- const q15_t *pTwiddleBReal; /**< points to the imag twiddle factor table. */
- const riscv_cfft_instance_q15 *pCfft; /**< points to the complex FFT instance. */
- } riscv_rfft_instance_q15;
-
- riscv_status riscv_rfft_init_q15(
- riscv_rfft_instance_q15 * S,
- uint32_t fftLenReal,
- uint32_t ifftFlagR,
- uint32_t bitReverseFlag);
-
- void riscv_rfft_q15(
- const riscv_rfft_instance_q15 * S,
- q15_t * pSrc,
- q15_t * pDst);
-
- /**
- * @brief Instance structure for the Q31 RFFT/RIFFT function.
- */
- typedef struct
- {
- uint32_t fftLenReal; /**< length of the real FFT. */
- uint8_t ifftFlagR; /**< flag that selects forward (ifftFlagR=0) or inverse (ifftFlagR=1) transform. */
- uint8_t bitReverseFlagR; /**< flag that enables (bitReverseFlagR=1) or disables (bitReverseFlagR=0) bit reversal of output. */
- uint32_t twidCoefRModifier; /**< twiddle coefficient modifier that supports different size FFTs with the same twiddle factor table. */
- const q31_t *pTwiddleAReal; /**< points to the real twiddle factor table. */
- const q31_t *pTwiddleBReal; /**< points to the imag twiddle factor table. */
- const riscv_cfft_instance_q31 *pCfft; /**< points to the complex FFT instance. */
- } riscv_rfft_instance_q31;
-
- riscv_status riscv_rfft_init_q31(
- riscv_rfft_instance_q31 * S,
- uint32_t fftLenReal,
- uint32_t ifftFlagR,
- uint32_t bitReverseFlag);
-
- void riscv_rfft_q31(
- const riscv_rfft_instance_q31 * S,
- q31_t * pSrc,
- q31_t * pDst);
-
- /**
- * @brief Instance structure for the floating-point RFFT/RIFFT function.
- */
- typedef struct
- {
- uint32_t fftLenReal; /**< length of the real FFT. */
- uint16_t fftLenBy2; /**< length of the complex FFT. */
- uint8_t ifftFlagR; /**< flag that selects forward (ifftFlagR=0) or inverse (ifftFlagR=1) transform. */
- uint8_t bitReverseFlagR; /**< flag that enables (bitReverseFlagR=1) or disables (bitReverseFlagR=0) bit reversal of output. */
- uint32_t twidCoefRModifier; /**< twiddle coefficient modifier that supports different size FFTs with the same twiddle factor table. */
- const float32_t *pTwiddleAReal; /**< points to the real twiddle factor table. */
- const float32_t *pTwiddleBReal; /**< points to the imag twiddle factor table. */
- riscv_cfft_radix4_instance_f32 *pCfft; /**< points to the complex FFT instance. */
- } riscv_rfft_instance_f32;
-
- riscv_status riscv_rfft_init_f32(
- riscv_rfft_instance_f32 * S,
- riscv_cfft_radix4_instance_f32 * S_CFFT,
- uint32_t fftLenReal,
- uint32_t ifftFlagR,
- uint32_t bitReverseFlag);
-
- void riscv_rfft_f32(
- const riscv_rfft_instance_f32 * S,
- float32_t * pSrc,
- float32_t * pDst);
-
- /**
- * @brief Instance structure for the floating-point RFFT/RIFFT function.
- */
-typedef struct
- {
- riscv_cfft_instance_f32 Sint; /**< Internal CFFT structure. */
- uint16_t fftLenRFFT; /**< length of the real sequence */
- const float32_t * pTwiddleRFFT; /**< Twiddle factors real stage */
- } riscv_rfft_fast_instance_f32 ;
-
-riscv_status riscv_rfft_fast_init_f32 (
- riscv_rfft_fast_instance_f32 * S,
- uint16_t fftLen);
-
-riscv_status riscv_rfft_32_fast_init_f32 ( riscv_rfft_fast_instance_f32 * S );
-
-riscv_status riscv_rfft_64_fast_init_f32 ( riscv_rfft_fast_instance_f32 * S );
-
-riscv_status riscv_rfft_128_fast_init_f32 ( riscv_rfft_fast_instance_f32 * S );
-
-riscv_status riscv_rfft_256_fast_init_f32 ( riscv_rfft_fast_instance_f32 * S );
-
-riscv_status riscv_rfft_512_fast_init_f32 ( riscv_rfft_fast_instance_f32 * S );
-
-riscv_status riscv_rfft_1024_fast_init_f32 ( riscv_rfft_fast_instance_f32 * S );
-
-riscv_status riscv_rfft_2048_fast_init_f32 ( riscv_rfft_fast_instance_f32 * S );
-
-riscv_status riscv_rfft_4096_fast_init_f32 ( riscv_rfft_fast_instance_f32 * S );
-
-
- void riscv_rfft_fast_f32(
- riscv_rfft_fast_instance_f32 * S,
- float32_t * p, float32_t * pOut,
- uint8_t ifftFlag);
-
- /**
- * @brief Instance structure for the floating-point DCT4/IDCT4 function.
- */
- typedef struct
- {
- uint16_t N; /**< length of the DCT4. */
- uint16_t Nby2; /**< half of the length of the DCT4. */
- float32_t normalize; /**< normalizing factor. */
- const float32_t *pTwiddle; /**< points to the twiddle factor table. */
- const float32_t *pCosFactor; /**< points to the cosFactor table. */
- riscv_rfft_instance_f32 *pRfft; /**< points to the real FFT instance. */
- riscv_cfft_radix4_instance_f32 *pCfft; /**< points to the complex FFT instance. */
- } riscv_dct4_instance_f32;
-
-
- /**
- * @brief Initialization function for the floating-point DCT4/IDCT4.
- * @param[in,out] S points to an instance of floating-point DCT4/IDCT4 structure.
- * @param[in] S_RFFT points to an instance of floating-point RFFT/RIFFT structure.
- * @param[in] S_CFFT points to an instance of floating-point CFFT/CIFFT structure.
- * @param[in] N length of the DCT4.
- * @param[in] Nby2 half of the length of the DCT4.
- * @param[in] normalize normalizing factor.
- * @return riscv_status function returns RISCV_MATH_SUCCESS if initialization is successful or RISCV_MATH_ARGUMENT_ERROR if fftLenReal is not a supported transform length.
- */
- riscv_status riscv_dct4_init_f32(
- riscv_dct4_instance_f32 * S,
- riscv_rfft_instance_f32 * S_RFFT,
- riscv_cfft_radix4_instance_f32 * S_CFFT,
- uint16_t N,
- uint16_t Nby2,
- float32_t normalize);
-
-
- /**
- * @brief Processing function for the floating-point DCT4/IDCT4.
- * @param[in] S points to an instance of the floating-point DCT4/IDCT4 structure.
- * @param[in] pState points to state buffer.
- * @param[in,out] pInlineBuffer points to the in-place input and output buffer.
- */
- void riscv_dct4_f32(
- const riscv_dct4_instance_f32 * S,
- float32_t * pState,
- float32_t * pInlineBuffer);
-
-
- /**
- * @brief Instance structure for the Q31 DCT4/IDCT4 function.
- */
- typedef struct
- {
- uint16_t N; /**< length of the DCT4. */
- uint16_t Nby2; /**< half of the length of the DCT4. */
- q31_t normalize; /**< normalizing factor. */
- const q31_t *pTwiddle; /**< points to the twiddle factor table. */
- const q31_t *pCosFactor; /**< points to the cosFactor table. */
- riscv_rfft_instance_q31 *pRfft; /**< points to the real FFT instance. */
- riscv_cfft_radix4_instance_q31 *pCfft; /**< points to the complex FFT instance. */
- } riscv_dct4_instance_q31;
-
-
- /**
- * @brief Initialization function for the Q31 DCT4/IDCT4.
- * @param[in,out] S points to an instance of Q31 DCT4/IDCT4 structure.
- * @param[in] S_RFFT points to an instance of Q31 RFFT/RIFFT structure
- * @param[in] S_CFFT points to an instance of Q31 CFFT/CIFFT structure
- * @param[in] N length of the DCT4.
- * @param[in] Nby2 half of the length of the DCT4.
- * @param[in] normalize normalizing factor.
- * @return riscv_status function returns RISCV_MATH_SUCCESS if initialization is successful or RISCV_MATH_ARGUMENT_ERROR if N is not a supported transform length.
- */
- riscv_status riscv_dct4_init_q31(
- riscv_dct4_instance_q31 * S,
- riscv_rfft_instance_q31 * S_RFFT,
- riscv_cfft_radix4_instance_q31 * S_CFFT,
- uint16_t N,
- uint16_t Nby2,
- q31_t normalize);
-
-
- /**
- * @brief Processing function for the Q31 DCT4/IDCT4.
- * @param[in] S points to an instance of the Q31 DCT4 structure.
- * @param[in] pState points to state buffer.
- * @param[in,out] pInlineBuffer points to the in-place input and output buffer.
- */
- void riscv_dct4_q31(
- const riscv_dct4_instance_q31 * S,
- q31_t * pState,
- q31_t * pInlineBuffer);
-
-
- /**
- * @brief Instance structure for the Q15 DCT4/IDCT4 function.
- */
- typedef struct
- {
- uint16_t N; /**< length of the DCT4. */
- uint16_t Nby2; /**< half of the length of the DCT4. */
- q15_t normalize; /**< normalizing factor. */
- const q15_t *pTwiddle; /**< points to the twiddle factor table. */
- const q15_t *pCosFactor; /**< points to the cosFactor table. */
- riscv_rfft_instance_q15 *pRfft; /**< points to the real FFT instance. */
- riscv_cfft_radix4_instance_q15 *pCfft; /**< points to the complex FFT instance. */
- } riscv_dct4_instance_q15;
-
-
- /**
- * @brief Initialization function for the Q15 DCT4/IDCT4.
- * @param[in,out] S points to an instance of Q15 DCT4/IDCT4 structure.
- * @param[in] S_RFFT points to an instance of Q15 RFFT/RIFFT structure.
- * @param[in] S_CFFT points to an instance of Q15 CFFT/CIFFT structure.
- * @param[in] N length of the DCT4.
- * @param[in] Nby2 half of the length of the DCT4.
- * @param[in] normalize normalizing factor.
- * @return riscv_status function returns RISCV_MATH_SUCCESS if initialization is successful or RISCV_MATH_ARGUMENT_ERROR if N is not a supported transform length.
- */
- riscv_status riscv_dct4_init_q15(
- riscv_dct4_instance_q15 * S,
- riscv_rfft_instance_q15 * S_RFFT,
- riscv_cfft_radix4_instance_q15 * S_CFFT,
- uint16_t N,
- uint16_t Nby2,
- q15_t normalize);
-
-
- /**
- * @brief Processing function for the Q15 DCT4/IDCT4.
- * @param[in] S points to an instance of the Q15 DCT4 structure.
- * @param[in] pState points to state buffer.
- * @param[in,out] pInlineBuffer points to the in-place input and output buffer.
- */
- void riscv_dct4_q15(
- const riscv_dct4_instance_q15 * S,
- q15_t * pState,
- q15_t * pInlineBuffer);
-
-
- /**
- * @brief Floating-point vector addition.
- * @param[in] pSrcA points to the first input vector
- * @param[in] pSrcB points to the second input vector
- * @param[out] pDst points to the output vector
- * @param[in] blockSize number of samples in each vector
- */
- void riscv_add_f32(
- const float32_t * pSrcA,
- const float32_t * pSrcB,
- float32_t * pDst,
- uint32_t blockSize);
-
-
- /**
- * @brief Q7 vector addition.
- * @param[in] pSrcA points to the first input vector
- * @param[in] pSrcB points to the second input vector
- * @param[out] pDst points to the output vector
- * @param[in] blockSize number of samples in each vector
- */
- void riscv_add_q7(
- const q7_t * pSrcA,
- const q7_t * pSrcB,
- q7_t * pDst,
- uint32_t blockSize);
-
-
- /**
- * @brief Q15 vector addition.
- * @param[in] pSrcA points to the first input vector
- * @param[in] pSrcB points to the second input vector
- * @param[out] pDst points to the output vector
- * @param[in] blockSize number of samples in each vector
- */
- void riscv_add_q15(
- const q15_t * pSrcA,
- const q15_t * pSrcB,
- q15_t * pDst,
- uint32_t blockSize);
-
-
- /**
- * @brief Q31 vector addition.
- * @param[in] pSrcA points to the first input vector
- * @param[in] pSrcB points to the second input vector
- * @param[out] pDst points to the output vector
- * @param[in] blockSize number of samples in each vector
- */
- void riscv_add_q31(
- const q31_t * pSrcA,
- const q31_t * pSrcB,
- q31_t * pDst,
- uint32_t blockSize);
-
-
- /**
- * @brief Floating-point vector subtraction.
- * @param[in] pSrcA points to the first input vector
- * @param[in] pSrcB points to the second input vector
- * @param[out] pDst points to the output vector
- * @param[in] blockSize number of samples in each vector
- */
- void riscv_sub_f32(
- const float32_t * pSrcA,
- const float32_t * pSrcB,
- float32_t * pDst,
- uint32_t blockSize);
-
-
- /**
- * @brief Q7 vector subtraction.
- * @param[in] pSrcA points to the first input vector
- * @param[in] pSrcB points to the second input vector
- * @param[out] pDst points to the output vector
- * @param[in] blockSize number of samples in each vector
- */
- void riscv_sub_q7(
- const q7_t * pSrcA,
- const q7_t * pSrcB,
- q7_t * pDst,
- uint32_t blockSize);
-
-
- /**
- * @brief Q15 vector subtraction.
- * @param[in] pSrcA points to the first input vector
- * @param[in] pSrcB points to the second input vector
- * @param[out] pDst points to the output vector
- * @param[in] blockSize number of samples in each vector
- */
- void riscv_sub_q15(
- const q15_t * pSrcA,
- const q15_t * pSrcB,
- q15_t * pDst,
- uint32_t blockSize);
-
-
- /**
- * @brief Q31 vector subtraction.
- * @param[in] pSrcA points to the first input vector
- * @param[in] pSrcB points to the second input vector
- * @param[out] pDst points to the output vector
- * @param[in] blockSize number of samples in each vector
- */
- void riscv_sub_q31(
- const q31_t * pSrcA,
- const q31_t * pSrcB,
- q31_t * pDst,
- uint32_t blockSize);
-
-
- /**
- * @brief Multiplies a floating-point vector by a scalar.
- * @param[in] pSrc points to the input vector
- * @param[in] scale scale factor to be applied
- * @param[out] pDst points to the output vector
- * @param[in] blockSize number of samples in the vector
- */
- void riscv_scale_f32(
- const float32_t * pSrc,
- float32_t scale,
- float32_t * pDst,
- uint32_t blockSize);
-
-
- /**
- * @brief Multiplies a Q7 vector by a scalar.
- * @param[in] pSrc points to the input vector
- * @param[in] scaleFract fractional portion of the scale value
- * @param[in] shift number of bits to shift the result by
- * @param[out] pDst points to the output vector
- * @param[in] blockSize number of samples in the vector
- */
- void riscv_scale_q7(
- const q7_t * pSrc,
- q7_t scaleFract,
- int8_t shift,
- q7_t * pDst,
- uint32_t blockSize);
-
-
- /**
- * @brief Multiplies a Q15 vector by a scalar.
- * @param[in] pSrc points to the input vector
- * @param[in] scaleFract fractional portion of the scale value
- * @param[in] shift number of bits to shift the result by
- * @param[out] pDst points to the output vector
- * @param[in] blockSize number of samples in the vector
- */
- void riscv_scale_q15(
- const q15_t * pSrc,
- q15_t scaleFract,
- int8_t shift,
- q15_t * pDst,
- uint32_t blockSize);
-
-
- /**
- * @brief Multiplies a Q31 vector by a scalar.
- * @param[in] pSrc points to the input vector
- * @param[in] scaleFract fractional portion of the scale value
- * @param[in] shift number of bits to shift the result by
- * @param[out] pDst points to the output vector
- * @param[in] blockSize number of samples in the vector
- */
- void riscv_scale_q31(
- const q31_t * pSrc,
- q31_t scaleFract,
- int8_t shift,
- q31_t * pDst,
- uint32_t blockSize);
-
-
- /**
- * @brief Q7 vector absolute value.
- * @param[in] pSrc points to the input buffer
- * @param[out] pDst points to the output buffer
- * @param[in] blockSize number of samples in each vector
- */
- void riscv_abs_q7(
- const q7_t * pSrc,
- q7_t * pDst,
- uint32_t blockSize);
-
-
- /**
- * @brief Floating-point vector absolute value.
- * @param[in] pSrc points to the input buffer
- * @param[out] pDst points to the output buffer
- * @param[in] blockSize number of samples in each vector
- */
- void riscv_abs_f32(
- const float32_t * pSrc,
- float32_t * pDst,
- uint32_t blockSize);
-
-
- /**
- * @brief Q15 vector absolute value.
- * @param[in] pSrc points to the input buffer
- * @param[out] pDst points to the output buffer
- * @param[in] blockSize number of samples in each vector
- */
- void riscv_abs_q15(
- const q15_t * pSrc,
- q15_t * pDst,
- uint32_t blockSize);
-
-
- /**
- * @brief Q31 vector absolute value.
- * @param[in] pSrc points to the input buffer
- * @param[out] pDst points to the output buffer
- * @param[in] blockSize number of samples in each vector
- */
- void riscv_abs_q31(
- const q31_t * pSrc,
- q31_t * pDst,
- uint32_t blockSize);
-
-
- /**
- * @brief Dot product of floating-point vectors.
- * @param[in] pSrcA points to the first input vector
- * @param[in] pSrcB points to the second input vector
- * @param[in] blockSize number of samples in each vector
- * @param[out] result output result returned here
- */
- void riscv_dot_prod_f32(
- const float32_t * pSrcA,
- const float32_t * pSrcB,
- uint32_t blockSize,
- float32_t * result);
-
-
- /**
- * @brief Dot product of Q7 vectors.
- * @param[in] pSrcA points to the first input vector
- * @param[in] pSrcB points to the second input vector
- * @param[in] blockSize number of samples in each vector
- * @param[out] result output result returned here
- */
- void riscv_dot_prod_q7(
- const q7_t * pSrcA,
- const q7_t * pSrcB,
- uint32_t blockSize,
- q31_t * result);
-
-
- /**
- * @brief Dot product of Q15 vectors.
- * @param[in] pSrcA points to the first input vector
- * @param[in] pSrcB points to the second input vector
- * @param[in] blockSize number of samples in each vector
- * @param[out] result output result returned here
- */
- void riscv_dot_prod_q15(
- const q15_t * pSrcA,
- const q15_t * pSrcB,
- uint32_t blockSize,
- q63_t * result);
-
-
- /**
- * @brief Dot product of Q31 vectors.
- * @param[in] pSrcA points to the first input vector
- * @param[in] pSrcB points to the second input vector
- * @param[in] blockSize number of samples in each vector
- * @param[out] result output result returned here
- */
- void riscv_dot_prod_q31(
- const q31_t * pSrcA,
- const q31_t * pSrcB,
- uint32_t blockSize,
- q63_t * result);
-
-
- /**
- * @brief Shifts the elements of a Q7 vector a specified number of bits.
- * @param[in] pSrc points to the input vector
- * @param[in] shiftBits number of bits to shift. A positive value shifts left; a negative value shifts right.
- * @param[out] pDst points to the output vector
- * @param[in] blockSize number of samples in the vector
- */
- void riscv_shift_q7(
- const q7_t * pSrc,
- int8_t shiftBits,
- q7_t * pDst,
- uint32_t blockSize);
-
-
- /**
- * @brief Shifts the elements of a Q15 vector a specified number of bits.
- * @param[in] pSrc points to the input vector
- * @param[in] shiftBits number of bits to shift. A positive value shifts left; a negative value shifts right.
- * @param[out] pDst points to the output vector
- * @param[in] blockSize number of samples in the vector
- */
- void riscv_shift_q15(
- const q15_t * pSrc,
- int8_t shiftBits,
- q15_t * pDst,
- uint32_t blockSize);
-
-
- /**
- * @brief Shifts the elements of a Q31 vector a specified number of bits.
- * @param[in] pSrc points to the input vector
- * @param[in] shiftBits number of bits to shift. A positive value shifts left; a negative value shifts right.
- * @param[out] pDst points to the output vector
- * @param[in] blockSize number of samples in the vector
- */
- void riscv_shift_q31(
- const q31_t * pSrc,
- int8_t shiftBits,
- q31_t * pDst,
- uint32_t blockSize);
-
-
- /**
- * @brief Adds a constant offset to a floating-point vector.
- * @param[in] pSrc points to the input vector
- * @param[in] offset is the offset to be added
- * @param[out] pDst points to the output vector
- * @param[in] blockSize number of samples in the vector
- */
- void riscv_offset_f32(
- const float32_t * pSrc,
- float32_t offset,
- float32_t * pDst,
- uint32_t blockSize);
-
-
- /**
- * @brief Adds a constant offset to a Q7 vector.
- * @param[in] pSrc points to the input vector
- * @param[in] offset is the offset to be added
- * @param[out] pDst points to the output vector
- * @param[in] blockSize number of samples in the vector
- */
- void riscv_offset_q7(
- const q7_t * pSrc,
- q7_t offset,
- q7_t * pDst,
- uint32_t blockSize);
-
-
- /**
- * @brief Adds a constant offset to a Q15 vector.
- * @param[in] pSrc points to the input vector
- * @param[in] offset is the offset to be added
- * @param[out] pDst points to the output vector
- * @param[in] blockSize number of samples in the vector
- */
- void riscv_offset_q15(
- const q15_t * pSrc,
- q15_t offset,
- q15_t * pDst,
- uint32_t blockSize);
-
-
- /**
- * @brief Adds a constant offset to a Q31 vector.
- * @param[in] pSrc points to the input vector
- * @param[in] offset is the offset to be added
- * @param[out] pDst points to the output vector
- * @param[in] blockSize number of samples in the vector
- */
- void riscv_offset_q31(
- const q31_t * pSrc,
- q31_t offset,
- q31_t * pDst,
- uint32_t blockSize);
-
-
- /**
- * @brief Negates the elements of a floating-point vector.
- * @param[in] pSrc points to the input vector
- * @param[out] pDst points to the output vector
- * @param[in] blockSize number of samples in the vector
- */
- void riscv_negate_f32(
- const float32_t * pSrc,
- float32_t * pDst,
- uint32_t blockSize);
-
-
- /**
- * @brief Negates the elements of a Q7 vector.
- * @param[in] pSrc points to the input vector
- * @param[out] pDst points to the output vector
- * @param[in] blockSize number of samples in the vector
- */
- void riscv_negate_q7(
- const q7_t * pSrc,
- q7_t * pDst,
- uint32_t blockSize);
-
-
- /**
- * @brief Negates the elements of a Q15 vector.
- * @param[in] pSrc points to the input vector
- * @param[out] pDst points to the output vector
- * @param[in] blockSize number of samples in the vector
- */
- void riscv_negate_q15(
- const q15_t * pSrc,
- q15_t * pDst,
- uint32_t blockSize);
-
-
- /**
- * @brief Negates the elements of a Q31 vector.
- * @param[in] pSrc points to the input vector
- * @param[out] pDst points to the output vector
- * @param[in] blockSize number of samples in the vector
- */
- void riscv_negate_q31(
- const q31_t * pSrc,
- q31_t * pDst,
- uint32_t blockSize);
-
-
- /**
- * @brief Copies the elements of a floating-point vector.
- * @param[in] pSrc input pointer
- * @param[out] pDst output pointer
- * @param[in] blockSize number of samples to process
- */
- void riscv_copy_f32(
- const float32_t * pSrc,
- float32_t * pDst,
- uint32_t blockSize);
-
-
- /**
- * @brief Copies the elements of a Q7 vector.
- * @param[in] pSrc input pointer
- * @param[out] pDst output pointer
- * @param[in] blockSize number of samples to process
- */
- void riscv_copy_q7(
- const q7_t * pSrc,
- q7_t * pDst,
- uint32_t blockSize);
-
-
- /**
- * @brief Copies the elements of a Q15 vector.
- * @param[in] pSrc input pointer
- * @param[out] pDst output pointer
- * @param[in] blockSize number of samples to process
- */
- void riscv_copy_q15(
- const q15_t * pSrc,
- q15_t * pDst,
- uint32_t blockSize);
-
-
- /**
- * @brief Copies the elements of a Q31 vector.
- * @param[in] pSrc input pointer
- * @param[out] pDst output pointer
- * @param[in] blockSize number of samples to process
- */
- void riscv_copy_q31(
- const q31_t * pSrc,
- q31_t * pDst,
- uint32_t blockSize);
-
-
- /**
- * @brief Fills a constant value into a floating-point vector.
- * @param[in] value input value to be filled
- * @param[out] pDst output pointer
- * @param[in] blockSize number of samples to process
- */
- void riscv_fill_f32(
- float32_t value,
- float32_t * pDst,
- uint32_t blockSize);
-
-
- /**
- * @brief Fills a constant value into a Q7 vector.
- * @param[in] value input value to be filled
- * @param[out] pDst output pointer
- * @param[in] blockSize number of samples to process
- */
- void riscv_fill_q7(
- q7_t value,
- q7_t * pDst,
- uint32_t blockSize);
-
-
- /**
- * @brief Fills a constant value into a Q15 vector.
- * @param[in] value input value to be filled
- * @param[out] pDst output pointer
- * @param[in] blockSize number of samples to process
- */
- void riscv_fill_q15(
- q15_t value,
- q15_t * pDst,
- uint32_t blockSize);
-
-
- /**
- * @brief Fills a constant value into a Q31 vector.
- * @param[in] value input value to be filled
- * @param[out] pDst output pointer
- * @param[in] blockSize number of samples to process
- */
- void riscv_fill_q31(
- q31_t value,
- q31_t * pDst,
- uint32_t blockSize);
-
-
-/**
- * @brief Convolution of floating-point sequences.
- * @param[in] pSrcA points to the first input sequence.
- * @param[in] srcALen length of the first input sequence.
- * @param[in] pSrcB points to the second input sequence.
- * @param[in] srcBLen length of the second input sequence.
- * @param[out] pDst points to the location where the output result is written. Length srcALen+srcBLen-1.
- */
- void riscv_conv_f32(
- const float32_t * pSrcA,
- uint32_t srcALen,
- const float32_t * pSrcB,
- uint32_t srcBLen,
- float32_t * pDst);
-
-
- /**
- * @brief Convolution of Q15 sequences.
- * @param[in] pSrcA points to the first input sequence.
- * @param[in] srcALen length of the first input sequence.
- * @param[in] pSrcB points to the second input sequence.
- * @param[in] srcBLen length of the second input sequence.
- * @param[out] pDst points to the block of output data Length srcALen+srcBLen-1.
- * @param[in] pScratch1 points to scratch buffer of size max(srcALen, srcBLen) + 2*min(srcALen, srcBLen) - 2.
- * @param[in] pScratch2 points to scratch buffer of size min(srcALen, srcBLen).
- */
- void riscv_conv_opt_q15(
- const q15_t * pSrcA,
- uint32_t srcALen,
- const q15_t * pSrcB,
- uint32_t srcBLen,
- q15_t * pDst,
- q15_t * pScratch1,
- q15_t * pScratch2);
-
-
-/**
- * @brief Convolution of Q15 sequences.
- * @param[in] pSrcA points to the first input sequence.
- * @param[in] srcALen length of the first input sequence.
- * @param[in] pSrcB points to the second input sequence.
- * @param[in] srcBLen length of the second input sequence.
- * @param[out] pDst points to the location where the output result is written. Length srcALen+srcBLen-1.
- */
- void riscv_conv_q15(
- const q15_t * pSrcA,
- uint32_t srcALen,
- const q15_t * pSrcB,
- uint32_t srcBLen,
- q15_t * pDst);
-
-
- /**
- * @brief Convolution of Q15 sequences (fast version) for RISC-V Core with DSP enabled
- * @param[in] pSrcA points to the first input sequence.
- * @param[in] srcALen length of the first input sequence.
- * @param[in] pSrcB points to the second input sequence.
- * @param[in] srcBLen length of the second input sequence.
- * @param[out] pDst points to the block of output data Length srcALen+srcBLen-1.
- */
- void riscv_conv_fast_q15(
- const q15_t * pSrcA,
- uint32_t srcALen,
- const q15_t * pSrcB,
- uint32_t srcBLen,
- q15_t * pDst);
-
-
- /**
- * @brief Convolution of Q15 sequences (fast version) for RISC-V Core with DSP enabled
- * @param[in] pSrcA points to the first input sequence.
- * @param[in] srcALen length of the first input sequence.
- * @param[in] pSrcB points to the second input sequence.
- * @param[in] srcBLen length of the second input sequence.
- * @param[out] pDst points to the block of output data Length srcALen+srcBLen-1.
- * @param[in] pScratch1 points to scratch buffer of size max(srcALen, srcBLen) + 2*min(srcALen, srcBLen) - 2.
- * @param[in] pScratch2 points to scratch buffer of size min(srcALen, srcBLen).
- */
- void riscv_conv_fast_opt_q15(
- const q15_t * pSrcA,
- uint32_t srcALen,
- const q15_t * pSrcB,
- uint32_t srcBLen,
- q15_t * pDst,
- q15_t * pScratch1,
- q15_t * pScratch2);
-
-
- /**
- * @brief Convolution of Q31 sequences.
- * @param[in] pSrcA points to the first input sequence.
- * @param[in] srcALen length of the first input sequence.
- * @param[in] pSrcB points to the second input sequence.
- * @param[in] srcBLen length of the second input sequence.
- * @param[out] pDst points to the block of output data Length srcALen+srcBLen-1.
- */
- void riscv_conv_q31(
- const q31_t * pSrcA,
- uint32_t srcALen,
- const q31_t * pSrcB,
- uint32_t srcBLen,
- q31_t * pDst);
-
-
- /**
- * @brief Convolution of Q31 sequences (fast version) for RISC-V Core with DSP enabled
- * @param[in] pSrcA points to the first input sequence.
- * @param[in] srcALen length of the first input sequence.
- * @param[in] pSrcB points to the second input sequence.
- * @param[in] srcBLen length of the second input sequence.
- * @param[out] pDst points to the block of output data Length srcALen+srcBLen-1.
- */
- void riscv_conv_fast_q31(
- const q31_t * pSrcA,
- uint32_t srcALen,
- const q31_t * pSrcB,
- uint32_t srcBLen,
- q31_t * pDst);
-
-
- /**
- * @brief Convolution of Q7 sequences.
- * @param[in] pSrcA points to the first input sequence.
- * @param[in] srcALen length of the first input sequence.
- * @param[in] pSrcB points to the second input sequence.
- * @param[in] srcBLen length of the second input sequence.
- * @param[out] pDst points to the block of output data Length srcALen+srcBLen-1.
- * @param[in] pScratch1 points to scratch buffer(of type q15_t) of size max(srcALen, srcBLen) + 2*min(srcALen, srcBLen) - 2.
- * @param[in] pScratch2 points to scratch buffer (of type q15_t) of size min(srcALen, srcBLen).
- */
- void riscv_conv_opt_q7(
- const q7_t * pSrcA,
- uint32_t srcALen,
- const q7_t * pSrcB,
- uint32_t srcBLen,
- q7_t * pDst,
- q15_t * pScratch1,
- q15_t * pScratch2);
-
-
- /**
- * @brief Convolution of Q7 sequences.
- * @param[in] pSrcA points to the first input sequence.
- * @param[in] srcALen length of the first input sequence.
- * @param[in] pSrcB points to the second input sequence.
- * @param[in] srcBLen length of the second input sequence.
- * @param[out] pDst points to the block of output data Length srcALen+srcBLen-1.
- */
- void riscv_conv_q7(
- const q7_t * pSrcA,
- uint32_t srcALen,
- const q7_t * pSrcB,
- uint32_t srcBLen,
- q7_t * pDst);
-
-
- /**
- * @brief Partial convolution of floating-point sequences.
- * @param[in] pSrcA points to the first input sequence.
- * @param[in] srcALen length of the first input sequence.
- * @param[in] pSrcB points to the second input sequence.
- * @param[in] srcBLen length of the second input sequence.
- * @param[out] pDst points to the block of output data
- * @param[in] firstIndex is the first output sample to start with.
- * @param[in] numPoints is the number of output points to be computed.
- * @return Returns either RISCV_MATH_SUCCESS if the function completed correctly or RISCV_MATH_ARGUMENT_ERROR if the requested subset is not in the range [0 srcALen+srcBLen-2].
- */
- riscv_status riscv_conv_partial_f32(
- const float32_t * pSrcA,
- uint32_t srcALen,
- const float32_t * pSrcB,
- uint32_t srcBLen,
- float32_t * pDst,
- uint32_t firstIndex,
- uint32_t numPoints);
-
-
- /**
- * @brief Partial convolution of Q15 sequences.
- * @param[in] pSrcA points to the first input sequence.
- * @param[in] srcALen length of the first input sequence.
- * @param[in] pSrcB points to the second input sequence.
- * @param[in] srcBLen length of the second input sequence.
- * @param[out] pDst points to the block of output data
- * @param[in] firstIndex is the first output sample to start with.
- * @param[in] numPoints is the number of output points to be computed.
- * @param[in] pScratch1 points to scratch buffer of size max(srcALen, srcBLen) + 2*min(srcALen, srcBLen) - 2.
- * @param[in] pScratch2 points to scratch buffer of size min(srcALen, srcBLen).
- * @return Returns either RISCV_MATH_SUCCESS if the function completed correctly or RISCV_MATH_ARGUMENT_ERROR if the requested subset is not in the range [0 srcALen+srcBLen-2].
- */
- riscv_status riscv_conv_partial_opt_q15(
- const q15_t * pSrcA,
- uint32_t srcALen,
- const q15_t * pSrcB,
- uint32_t srcBLen,
- q15_t * pDst,
- uint32_t firstIndex,
- uint32_t numPoints,
- q15_t * pScratch1,
- q15_t * pScratch2);
-
-
- /**
- * @brief Partial convolution of Q15 sequences.
- * @param[in] pSrcA points to the first input sequence.
- * @param[in] srcALen length of the first input sequence.
- * @param[in] pSrcB points to the second input sequence.
- * @param[in] srcBLen length of the second input sequence.
- * @param[out] pDst points to the block of output data
- * @param[in] firstIndex is the first output sample to start with.
- * @param[in] numPoints is the number of output points to be computed.
- * @return Returns either RISCV_MATH_SUCCESS if the function completed correctly or RISCV_MATH_ARGUMENT_ERROR if the requested subset is not in the range [0 srcALen+srcBLen-2].
- */
- riscv_status riscv_conv_partial_q15(
- const q15_t * pSrcA,
- uint32_t srcALen,
- const q15_t * pSrcB,
- uint32_t srcBLen,
- q15_t * pDst,
- uint32_t firstIndex,
- uint32_t numPoints);
-
-
- /**
- * @brief Partial convolution of Q15 sequences (fast version) for RISC-V Core with DSP enabled
- * @param[in] pSrcA points to the first input sequence.
- * @param[in] srcALen length of the first input sequence.
- * @param[in] pSrcB points to the second input sequence.
- * @param[in] srcBLen length of the second input sequence.
- * @param[out] pDst points to the block of output data
- * @param[in] firstIndex is the first output sample to start with.
- * @param[in] numPoints is the number of output points to be computed.
- * @return Returns either RISCV_MATH_SUCCESS if the function completed correctly or RISCV_MATH_ARGUMENT_ERROR if the requested subset is not in the range [0 srcALen+srcBLen-2].
- */
- riscv_status riscv_conv_partial_fast_q15(
- const q15_t * pSrcA,
- uint32_t srcALen,
- const q15_t * pSrcB,
- uint32_t srcBLen,
- q15_t * pDst,
- uint32_t firstIndex,
- uint32_t numPoints);
-
-
- /**
- * @brief Partial convolution of Q15 sequences (fast version) for RISC-V Core with DSP enabled
- * @param[in] pSrcA points to the first input sequence.
- * @param[in] srcALen length of the first input sequence.
- * @param[in] pSrcB points to the second input sequence.
- * @param[in] srcBLen length of the second input sequence.
- * @param[out] pDst points to the block of output data
- * @param[in] firstIndex is the first output sample to start with.
- * @param[in] numPoints is the number of output points to be computed.
- * @param[in] pScratch1 points to scratch buffer of size max(srcALen, srcBLen) + 2*min(srcALen, srcBLen) - 2.
- * @param[in] pScratch2 points to scratch buffer of size min(srcALen, srcBLen).
- * @return Returns either RISCV_MATH_SUCCESS if the function completed correctly or RISCV_MATH_ARGUMENT_ERROR if the requested subset is not in the range [0 srcALen+srcBLen-2].
- */
- riscv_status riscv_conv_partial_fast_opt_q15(
- const q15_t * pSrcA,
- uint32_t srcALen,
- const q15_t * pSrcB,
- uint32_t srcBLen,
- q15_t * pDst,
- uint32_t firstIndex,
- uint32_t numPoints,
- q15_t * pScratch1,
- q15_t * pScratch2);
-
-
- /**
- * @brief Partial convolution of Q31 sequences.
- * @param[in] pSrcA points to the first input sequence.
- * @param[in] srcALen length of the first input sequence.
- * @param[in] pSrcB points to the second input sequence.
- * @param[in] srcBLen length of the second input sequence.
- * @param[out] pDst points to the block of output data
- * @param[in] firstIndex is the first output sample to start with.
- * @param[in] numPoints is the number of output points to be computed.
- * @return Returns either RISCV_MATH_SUCCESS if the function completed correctly or RISCV_MATH_ARGUMENT_ERROR if the requested subset is not in the range [0 srcALen+srcBLen-2].
- */
- riscv_status riscv_conv_partial_q31(
- const q31_t * pSrcA,
- uint32_t srcALen,
- const q31_t * pSrcB,
- uint32_t srcBLen,
- q31_t * pDst,
- uint32_t firstIndex,
- uint32_t numPoints);
-
-
- /**
- * @brief Partial convolution of Q31 sequences (fast version) for RISC-V Core with DSP enabled
- * @param[in] pSrcA points to the first input sequence.
- * @param[in] srcALen length of the first input sequence.
- * @param[in] pSrcB points to the second input sequence.
- * @param[in] srcBLen length of the second input sequence.
- * @param[out] pDst points to the block of output data
- * @param[in] firstIndex is the first output sample to start with.
- * @param[in] numPoints is the number of output points to be computed.
- * @return Returns either RISCV_MATH_SUCCESS if the function completed correctly or RISCV_MATH_ARGUMENT_ERROR if the requested subset is not in the range [0 srcALen+srcBLen-2].
- */
- riscv_status riscv_conv_partial_fast_q31(
- const q31_t * pSrcA,
- uint32_t srcALen,
- const q31_t * pSrcB,
- uint32_t srcBLen,
- q31_t * pDst,
- uint32_t firstIndex,
- uint32_t numPoints);
-
-
- /**
- * @brief Partial convolution of Q7 sequences
- * @param[in] pSrcA points to the first input sequence.
- * @param[in] srcALen length of the first input sequence.
- * @param[in] pSrcB points to the second input sequence.
- * @param[in] srcBLen length of the second input sequence.
- * @param[out] pDst points to the block of output data
- * @param[in] firstIndex is the first output sample to start with.
- * @param[in] numPoints is the number of output points to be computed.
- * @param[in] pScratch1 points to scratch buffer(of type q15_t) of size max(srcALen, srcBLen) + 2*min(srcALen, srcBLen) - 2.
- * @param[in] pScratch2 points to scratch buffer (of type q15_t) of size min(srcALen, srcBLen).
- * @return Returns either RISCV_MATH_SUCCESS if the function completed correctly or RISCV_MATH_ARGUMENT_ERROR if the requested subset is not in the range [0 srcALen+srcBLen-2].
- */
- riscv_status riscv_conv_partial_opt_q7(
- const q7_t * pSrcA,
- uint32_t srcALen,
- const q7_t * pSrcB,
- uint32_t srcBLen,
- q7_t * pDst,
- uint32_t firstIndex,
- uint32_t numPoints,
- q15_t * pScratch1,
- q15_t * pScratch2);
-
-
-/**
- * @brief Partial convolution of Q7 sequences.
- * @param[in] pSrcA points to the first input sequence.
- * @param[in] srcALen length of the first input sequence.
- * @param[in] pSrcB points to the second input sequence.
- * @param[in] srcBLen length of the second input sequence.
- * @param[out] pDst points to the block of output data
- * @param[in] firstIndex is the first output sample to start with.
- * @param[in] numPoints is the number of output points to be computed.
- * @return Returns either RISCV_MATH_SUCCESS if the function completed correctly or RISCV_MATH_ARGUMENT_ERROR if the requested subset is not in the range [0 srcALen+srcBLen-2].
- */
- riscv_status riscv_conv_partial_q7(
- const q7_t * pSrcA,
- uint32_t srcALen,
- const q7_t * pSrcB,
- uint32_t srcBLen,
- q7_t * pDst,
- uint32_t firstIndex,
- uint32_t numPoints);
-
-
- /**
- * @brief Instance structure for the Q15 FIR decimator.
- */
- typedef struct
- {
- uint8_t M; /**< decimation factor. */
- uint16_t numTaps; /**< number of coefficients in the filter. */
- const q15_t *pCoeffs; /**< points to the coefficient array. The array is of length numTaps.*/
- q15_t *pState; /**< points to the state variable array. The array is of length numTaps+blockSize-1. */
- } riscv_fir_decimate_instance_q15;
-
- /**
- * @brief Instance structure for the Q31 FIR decimator.
- */
- typedef struct
- {
- uint8_t M; /**< decimation factor. */
- uint16_t numTaps; /**< number of coefficients in the filter. */
- const q31_t *pCoeffs; /**< points to the coefficient array. The array is of length numTaps.*/
- q31_t *pState; /**< points to the state variable array. The array is of length numTaps+blockSize-1. */
- } riscv_fir_decimate_instance_q31;
-
-/**
- @brief Instance structure for floating-point FIR decimator.
- */
-typedef struct
- {
- uint8_t M; /**< decimation factor. */
- uint16_t numTaps; /**< number of coefficients in the filter. */
- const float32_t *pCoeffs; /**< points to the coefficient array. The array is of length numTaps.*/
- float32_t *pState; /**< points to the state variable array. The array is of length numTaps+blockSize-1. */
- } riscv_fir_decimate_instance_f32;
-
-
-/**
- @brief Processing function for floating-point FIR decimator.
- @param[in] S points to an instance of the floating-point FIR decimator structure
- @param[in] pSrc points to the block of input data
- @param[out] pDst points to the block of output data
- @param[in] blockSize number of samples to process
- */
-void riscv_fir_decimate_f32(
- const riscv_fir_decimate_instance_f32 * S,
- const float32_t * pSrc,
- float32_t * pDst,
- uint32_t blockSize);
-
-
-/**
- @brief Initialization function for the floating-point FIR decimator.
- @param[in,out] S points to an instance of the floating-point FIR decimator structure
- @param[in] numTaps number of coefficients in the filter
- @param[in] M decimation factor
- @param[in] pCoeffs points to the filter coefficients
- @param[in] pState points to the state buffer
- @param[in] blockSize number of input samples to process per call
- @return execution status
- - \ref RISCV_MATH_SUCCESS : Operation successful
- - \ref RISCV_MATH_LENGTH_ERROR : blockSize is not a multiple of M
- */
-riscv_status riscv_fir_decimate_init_f32(
- riscv_fir_decimate_instance_f32 * S,
- uint16_t numTaps,
- uint8_t M,
- const float32_t * pCoeffs,
- float32_t * pState,
- uint32_t blockSize);
-
-
- /**
- * @brief Processing function for the Q15 FIR decimator.
- * @param[in] S points to an instance of the Q15 FIR decimator structure.
- * @param[in] pSrc points to the block of input data.
- * @param[out] pDst points to the block of output data
- * @param[in] blockSize number of input samples to process per call.
- */
- void riscv_fir_decimate_q15(
- const riscv_fir_decimate_instance_q15 * S,
- const q15_t * pSrc,
- q15_t * pDst,
- uint32_t blockSize);
-
-
- /**
- * @brief Processing function for the Q15 FIR decimator (fast variant) for RISC-V Core with DSP enabled.
- * @param[in] S points to an instance of the Q15 FIR decimator structure.
- * @param[in] pSrc points to the block of input data.
- * @param[out] pDst points to the block of output data
- * @param[in] blockSize number of input samples to process per call.
- */
- void riscv_fir_decimate_fast_q15(
- const riscv_fir_decimate_instance_q15 * S,
- const q15_t * pSrc,
- q15_t * pDst,
- uint32_t blockSize);
-
-
- /**
- * @brief Initialization function for the Q15 FIR decimator.
- * @param[in,out] S points to an instance of the Q15 FIR decimator structure.
- * @param[in] numTaps number of coefficients in the filter.
- * @param[in] M decimation factor.
- * @param[in] pCoeffs points to the filter coefficients.
- * @param[in] pState points to the state buffer.
- * @param[in] blockSize number of input samples to process per call.
- * @return The function returns RISCV_MATH_SUCCESS if initialization is successful or RISCV_MATH_LENGTH_ERROR if
- * blockSize is not a multiple of M.
- */
- riscv_status riscv_fir_decimate_init_q15(
- riscv_fir_decimate_instance_q15 * S,
- uint16_t numTaps,
- uint8_t M,
- const q15_t * pCoeffs,
- q15_t * pState,
- uint32_t blockSize);
-
-
- /**
- * @brief Processing function for the Q31 FIR decimator.
- * @param[in] S points to an instance of the Q31 FIR decimator structure.
- * @param[in] pSrc points to the block of input data.
- * @param[out] pDst points to the block of output data
- * @param[in] blockSize number of input samples to process per call.
- */
- void riscv_fir_decimate_q31(
- const riscv_fir_decimate_instance_q31 * S,
- const q31_t * pSrc,
- q31_t * pDst,
- uint32_t blockSize);
-
- /**
- * @brief Processing function for the Q31 FIR decimator (fast variant) for RISC-V Core with DSP enabled.
- * @param[in] S points to an instance of the Q31 FIR decimator structure.
- * @param[in] pSrc points to the block of input data.
- * @param[out] pDst points to the block of output data
- * @param[in] blockSize number of input samples to process per call.
- */
- void riscv_fir_decimate_fast_q31(
- const riscv_fir_decimate_instance_q31 * S,
- const q31_t * pSrc,
- q31_t * pDst,
- uint32_t blockSize);
-
-
- /**
- * @brief Initialization function for the Q31 FIR decimator.
- * @param[in,out] S points to an instance of the Q31 FIR decimator structure.
- * @param[in] numTaps number of coefficients in the filter.
- * @param[in] M decimation factor.
- * @param[in] pCoeffs points to the filter coefficients.
- * @param[in] pState points to the state buffer.
- * @param[in] blockSize number of input samples to process per call.
- * @return The function returns RISCV_MATH_SUCCESS if initialization is successful or RISCV_MATH_LENGTH_ERROR if
- * blockSize is not a multiple of M.
- */
- riscv_status riscv_fir_decimate_init_q31(
- riscv_fir_decimate_instance_q31 * S,
- uint16_t numTaps,
- uint8_t M,
- const q31_t * pCoeffs,
- q31_t * pState,
- uint32_t blockSize);
-
-
- /**
- * @brief Instance structure for the Q15 FIR interpolator.
- */
- typedef struct
- {
- uint8_t L; /**< upsample factor. */
- uint16_t phaseLength; /**< length of each polyphase filter component. */
- const q15_t *pCoeffs; /**< points to the coefficient array. The array is of length L*phaseLength. */
- q15_t *pState; /**< points to the state variable array. The array is of length blockSize+phaseLength-1. */
- } riscv_fir_interpolate_instance_q15;
-
- /**
- * @brief Instance structure for the Q31 FIR interpolator.
- */
- typedef struct
- {
- uint8_t L; /**< upsample factor. */
- uint16_t phaseLength; /**< length of each polyphase filter component. */
- const q31_t *pCoeffs; /**< points to the coefficient array. The array is of length L*phaseLength. */
- q31_t *pState; /**< points to the state variable array. The array is of length blockSize+phaseLength-1. */
- } riscv_fir_interpolate_instance_q31;
-
- /**
- * @brief Instance structure for the floating-point FIR interpolator.
- */
- typedef struct
- {
- uint8_t L; /**< upsample factor. */
- uint16_t phaseLength; /**< length of each polyphase filter component. */
- const float32_t *pCoeffs; /**< points to the coefficient array. The array is of length L*phaseLength. */
- float32_t *pState; /**< points to the state variable array. The array is of length phaseLength+numTaps-1. */
- } riscv_fir_interpolate_instance_f32;
-
-
- /**
- * @brief Processing function for the Q15 FIR interpolator.
- * @param[in] S points to an instance of the Q15 FIR interpolator structure.
- * @param[in] pSrc points to the block of input data.
- * @param[out] pDst points to the block of output data.
- * @param[in] blockSize number of input samples to process per call.
- */
- void riscv_fir_interpolate_q15(
- const riscv_fir_interpolate_instance_q15 * S,
- const q15_t * pSrc,
- q15_t * pDst,
- uint32_t blockSize);
-
-
- /**
- * @brief Initialization function for the Q15 FIR interpolator.
- * @param[in,out] S points to an instance of the Q15 FIR interpolator structure.
- * @param[in] L upsample factor.
- * @param[in] numTaps number of filter coefficients in the filter.
- * @param[in] pCoeffs points to the filter coefficient buffer.
- * @param[in] pState points to the state buffer.
- * @param[in] blockSize number of input samples to process per call.
- * @return The function returns RISCV_MATH_SUCCESS if initialization is successful or RISCV_MATH_LENGTH_ERROR if
- * the filter length numTaps is not a multiple of the interpolation factor L.
- */
- riscv_status riscv_fir_interpolate_init_q15(
- riscv_fir_interpolate_instance_q15 * S,
- uint8_t L,
- uint16_t numTaps,
- const q15_t * pCoeffs,
- q15_t * pState,
- uint32_t blockSize);
-
-
- /**
- * @brief Processing function for the Q31 FIR interpolator.
- * @param[in] S points to an instance of the Q15 FIR interpolator structure.
- * @param[in] pSrc points to the block of input data.
- * @param[out] pDst points to the block of output data.
- * @param[in] blockSize number of input samples to process per call.
- */
- void riscv_fir_interpolate_q31(
- const riscv_fir_interpolate_instance_q31 * S,
- const q31_t * pSrc,
- q31_t * pDst,
- uint32_t blockSize);
-
-
- /**
- * @brief Initialization function for the Q31 FIR interpolator.
- * @param[in,out] S points to an instance of the Q31 FIR interpolator structure.
- * @param[in] L upsample factor.
- * @param[in] numTaps number of filter coefficients in the filter.
- * @param[in] pCoeffs points to the filter coefficient buffer.
- * @param[in] pState points to the state buffer.
- * @param[in] blockSize number of input samples to process per call.
- * @return The function returns RISCV_MATH_SUCCESS if initialization is successful or RISCV_MATH_LENGTH_ERROR if
- * the filter length numTaps is not a multiple of the interpolation factor L.
- */
- riscv_status riscv_fir_interpolate_init_q31(
- riscv_fir_interpolate_instance_q31 * S,
- uint8_t L,
- uint16_t numTaps,
- const q31_t * pCoeffs,
- q31_t * pState,
- uint32_t blockSize);
-
-
- /**
- * @brief Processing function for the floating-point FIR interpolator.
- * @param[in] S points to an instance of the floating-point FIR interpolator structure.
- * @param[in] pSrc points to the block of input data.
- * @param[out] pDst points to the block of output data.
- * @param[in] blockSize number of input samples to process per call.
- */
- void riscv_fir_interpolate_f32(
- const riscv_fir_interpolate_instance_f32 * S,
- const float32_t * pSrc,
- float32_t * pDst,
- uint32_t blockSize);
-
-
- /**
- * @brief Initialization function for the floating-point FIR interpolator.
- * @param[in,out] S points to an instance of the floating-point FIR interpolator structure.
- * @param[in] L upsample factor.
- * @param[in] numTaps number of filter coefficients in the filter.
- * @param[in] pCoeffs points to the filter coefficient buffer.
- * @param[in] pState points to the state buffer.
- * @param[in] blockSize number of input samples to process per call.
- * @return The function returns RISCV_MATH_SUCCESS if initialization is successful or RISCV_MATH_LENGTH_ERROR if
- * the filter length numTaps is not a multiple of the interpolation factor L.
- */
- riscv_status riscv_fir_interpolate_init_f32(
- riscv_fir_interpolate_instance_f32 * S,
- uint8_t L,
- uint16_t numTaps,
- const float32_t * pCoeffs,
- float32_t * pState,
- uint32_t blockSize);
-
-
- /**
- * @brief Instance structure for the high precision Q31 Biquad cascade filter.
- */
- typedef struct
- {
- uint8_t numStages; /**< number of 2nd order stages in the filter. Overall order is 2*numStages. */
- q63_t *pState; /**< points to the array of state coefficients. The array is of length 4*numStages. */
- const q31_t *pCoeffs; /**< points to the array of coefficients. The array is of length 5*numStages. */
- uint8_t postShift; /**< additional shift, in bits, applied to each output sample. */
- } riscv_biquad_cas_df1_32x64_ins_q31;
-
-
- /**
- * @param[in] S points to an instance of the high precision Q31 Biquad cascade filter structure.
- * @param[in] pSrc points to the block of input data.
- * @param[out] pDst points to the block of output data
- * @param[in] blockSize number of samples to process.
- */
- void riscv_biquad_cas_df1_32x64_q31(
- const riscv_biquad_cas_df1_32x64_ins_q31 * S,
- q31_t * pSrc,
- q31_t * pDst,
- uint32_t blockSize);
-
-
- /**
- * @param[in,out] S points to an instance of the high precision Q31 Biquad cascade filter structure.
- * @param[in] numStages number of 2nd order stages in the filter.
- * @param[in] pCoeffs points to the filter coefficients.
- * @param[in] pState points to the state buffer.
- * @param[in] postShift shift to be applied to the output. Varies according to the coefficients format
- */
- void riscv_biquad_cas_df1_32x64_init_q31(
- riscv_biquad_cas_df1_32x64_ins_q31 * S,
- uint8_t numStages,
- const q31_t * pCoeffs,
- q63_t * pState,
- uint8_t postShift);
-
-
- /**
- * @brief Instance structure for the floating-point transposed direct form II Biquad cascade filter.
- */
- typedef struct
- {
- uint8_t numStages; /**< number of 2nd order stages in the filter. Overall order is 2*numStages. */
- float32_t *pState; /**< points to the array of state coefficients. The array is of length 2*numStages. */
- const float32_t *pCoeffs; /**< points to the array of coefficients. The array is of length 5*numStages. */
- } riscv_biquad_cascade_df2T_instance_f32;
-
- /**
- * @brief Instance structure for the floating-point transposed direct form II Biquad cascade filter.
- */
- typedef struct
- {
- uint8_t numStages; /**< number of 2nd order stages in the filter. Overall order is 2*numStages. */
- float32_t *pState; /**< points to the array of state coefficients. The array is of length 4*numStages. */
- const float32_t *pCoeffs; /**< points to the array of coefficients. The array is of length 5*numStages. */
- } riscv_biquad_cascade_stereo_df2T_instance_f32;
-
- /**
- * @brief Instance structure for the floating-point transposed direct form II Biquad cascade filter.
- */
- typedef struct
- {
- uint8_t numStages; /**< number of 2nd order stages in the filter. Overall order is 2*numStages. */
- float64_t *pState; /**< points to the array of state coefficients. The array is of length 2*numStages. */
- float64_t *pCoeffs; /**< points to the array of coefficients. The array is of length 5*numStages. */
- } riscv_biquad_cascade_df2T_instance_f64;
-
-
- /**
- * @brief Processing function for the floating-point transposed direct form II Biquad cascade filter.
- * @param[in] S points to an instance of the filter data structure.
- * @param[in] pSrc points to the block of input data.
- * @param[out] pDst points to the block of output data
- * @param[in] blockSize number of samples to process.
- */
- void riscv_biquad_cascade_df2T_f32(
- const riscv_biquad_cascade_df2T_instance_f32 * S,
- const float32_t * pSrc,
- float32_t * pDst,
- uint32_t blockSize);
-
-
- /**
- * @brief Processing function for the floating-point transposed direct form II Biquad cascade filter. 2 channels
- * @param[in] S points to an instance of the filter data structure.
- * @param[in] pSrc points to the block of input data.
- * @param[out] pDst points to the block of output data
- * @param[in] blockSize number of samples to process.
- */
- void riscv_biquad_cascade_stereo_df2T_f32(
- const riscv_biquad_cascade_stereo_df2T_instance_f32 * S,
- const float32_t * pSrc,
- float32_t * pDst,
- uint32_t blockSize);
-
-
- /**
- * @brief Processing function for the floating-point transposed direct form II Biquad cascade filter.
- * @param[in] S points to an instance of the filter data structure.
- * @param[in] pSrc points to the block of input data.
- * @param[out] pDst points to the block of output data
- * @param[in] blockSize number of samples to process.
- */
- void riscv_biquad_cascade_df2T_f64(
- const riscv_biquad_cascade_df2T_instance_f64 * S,
- float64_t * pSrc,
- float64_t * pDst,
- uint32_t blockSize);
-
-
- /**
- * @brief Initialization function for the floating-point transposed direct form II Biquad cascade filter.
- * @param[in,out] S points to an instance of the filter data structure.
- * @param[in] numStages number of 2nd order stages in the filter.
- * @param[in] pCoeffs points to the filter coefficients.
- * @param[in] pState points to the state buffer.
- */
- void riscv_biquad_cascade_df2T_init_f32(
- riscv_biquad_cascade_df2T_instance_f32 * S,
- uint8_t numStages,
- const float32_t * pCoeffs,
- float32_t * pState);
-
-
- /**
- * @brief Initialization function for the floating-point transposed direct form II Biquad cascade filter.
- * @param[in,out] S points to an instance of the filter data structure.
- * @param[in] numStages number of 2nd order stages in the filter.
- * @param[in] pCoeffs points to the filter coefficients.
- * @param[in] pState points to the state buffer.
- */
- void riscv_biquad_cascade_stereo_df2T_init_f32(
- riscv_biquad_cascade_stereo_df2T_instance_f32 * S,
- uint8_t numStages,
- const float32_t * pCoeffs,
- float32_t * pState);
-
-
- /**
- * @brief Initialization function for the floating-point transposed direct form II Biquad cascade filter.
- * @param[in,out] S points to an instance of the filter data structure.
- * @param[in] numStages number of 2nd order stages in the filter.
- * @param[in] pCoeffs points to the filter coefficients.
- * @param[in] pState points to the state buffer.
- */
- void riscv_biquad_cascade_df2T_init_f64(
- riscv_biquad_cascade_df2T_instance_f64 * S,
- uint8_t numStages,
- float64_t * pCoeffs,
- float64_t * pState);
-
-
- /**
- * @brief Instance structure for the Q15 FIR lattice filter.
- */
- typedef struct
- {
- uint16_t numStages; /**< number of filter stages. */
- q15_t *pState; /**< points to the state variable array. The array is of length numStages. */
- const q15_t *pCoeffs; /**< points to the coefficient array. The array is of length numStages. */
- } riscv_fir_lattice_instance_q15;
-
- /**
- * @brief Instance structure for the Q31 FIR lattice filter.
- */
- typedef struct
- {
- uint16_t numStages; /**< number of filter stages. */
- q31_t *pState; /**< points to the state variable array. The array is of length numStages. */
- const q31_t *pCoeffs; /**< points to the coefficient array. The array is of length numStages. */
- } riscv_fir_lattice_instance_q31;
-
- /**
- * @brief Instance structure for the floating-point FIR lattice filter.
- */
- typedef struct
- {
- uint16_t numStages; /**< number of filter stages. */
- float32_t *pState; /**< points to the state variable array. The array is of length numStages. */
- const float32_t *pCoeffs; /**< points to the coefficient array. The array is of length numStages. */
- } riscv_fir_lattice_instance_f32;
-
-
- /**
- * @brief Initialization function for the Q15 FIR lattice filter.
- * @param[in] S points to an instance of the Q15 FIR lattice structure.
- * @param[in] numStages number of filter stages.
- * @param[in] pCoeffs points to the coefficient buffer. The array is of length numStages.
- * @param[in] pState points to the state buffer. The array is of length numStages.
- */
- void riscv_fir_lattice_init_q15(
- riscv_fir_lattice_instance_q15 * S,
- uint16_t numStages,
- const q15_t * pCoeffs,
- q15_t * pState);
-
-
- /**
- * @brief Processing function for the Q15 FIR lattice filter.
- * @param[in] S points to an instance of the Q15 FIR lattice structure.
- * @param[in] pSrc points to the block of input data.
- * @param[out] pDst points to the block of output data.
- * @param[in] blockSize number of samples to process.
- */
- void riscv_fir_lattice_q15(
- const riscv_fir_lattice_instance_q15 * S,
- const q15_t * pSrc,
- q15_t * pDst,
- uint32_t blockSize);
-
-
- /**
- * @brief Initialization function for the Q31 FIR lattice filter.
- * @param[in] S points to an instance of the Q31 FIR lattice structure.
- * @param[in] numStages number of filter stages.
- * @param[in] pCoeffs points to the coefficient buffer. The array is of length numStages.
- * @param[in] pState points to the state buffer. The array is of length numStages.
- */
- void riscv_fir_lattice_init_q31(
- riscv_fir_lattice_instance_q31 * S,
- uint16_t numStages,
- const q31_t * pCoeffs,
- q31_t * pState);
-
-
- /**
- * @brief Processing function for the Q31 FIR lattice filter.
- * @param[in] S points to an instance of the Q31 FIR lattice structure.
- * @param[in] pSrc points to the block of input data.
- * @param[out] pDst points to the block of output data
- * @param[in] blockSize number of samples to process.
- */
- void riscv_fir_lattice_q31(
- const riscv_fir_lattice_instance_q31 * S,
- const q31_t * pSrc,
- q31_t * pDst,
- uint32_t blockSize);
-
-
-/**
- * @brief Initialization function for the floating-point FIR lattice filter.
- * @param[in] S points to an instance of the floating-point FIR lattice structure.
- * @param[in] numStages number of filter stages.
- * @param[in] pCoeffs points to the coefficient buffer. The array is of length numStages.
- * @param[in] pState points to the state buffer. The array is of length numStages.
- */
- void riscv_fir_lattice_init_f32(
- riscv_fir_lattice_instance_f32 * S,
- uint16_t numStages,
- const float32_t * pCoeffs,
- float32_t * pState);
-
-
- /**
- * @brief Processing function for the floating-point FIR lattice filter.
- * @param[in] S points to an instance of the floating-point FIR lattice structure.
- * @param[in] pSrc points to the block of input data.
- * @param[out] pDst points to the block of output data
- * @param[in] blockSize number of samples to process.
- */
- void riscv_fir_lattice_f32(
- const riscv_fir_lattice_instance_f32 * S,
- const float32_t * pSrc,
- float32_t * pDst,
- uint32_t blockSize);
-
-
- /**
- * @brief Instance structure for the Q15 IIR lattice filter.
- */
- typedef struct
- {
- uint16_t numStages; /**< number of stages in the filter. */
- q15_t *pState; /**< points to the state variable array. The array is of length numStages+blockSize. */
- q15_t *pkCoeffs; /**< points to the reflection coefficient array. The array is of length numStages. */
- q15_t *pvCoeffs; /**< points to the ladder coefficient array. The array is of length numStages+1. */
- } riscv_iir_lattice_instance_q15;
-
- /**
- * @brief Instance structure for the Q31 IIR lattice filter.
- */
- typedef struct
- {
- uint16_t numStages; /**< number of stages in the filter. */
- q31_t *pState; /**< points to the state variable array. The array is of length numStages+blockSize. */
- q31_t *pkCoeffs; /**< points to the reflection coefficient array. The array is of length numStages. */
- q31_t *pvCoeffs; /**< points to the ladder coefficient array. The array is of length numStages+1. */
- } riscv_iir_lattice_instance_q31;
-
- /**
- * @brief Instance structure for the floating-point IIR lattice filter.
- */
- typedef struct
- {
- uint16_t numStages; /**< number of stages in the filter. */
- float32_t *pState; /**< points to the state variable array. The array is of length numStages+blockSize. */
- float32_t *pkCoeffs; /**< points to the reflection coefficient array. The array is of length numStages. */
- float32_t *pvCoeffs; /**< points to the ladder coefficient array. The array is of length numStages+1. */
- } riscv_iir_lattice_instance_f32;
-
-
- /**
- * @brief Processing function for the floating-point IIR lattice filter.
- * @param[in] S points to an instance of the floating-point IIR lattice structure.
- * @param[in] pSrc points to the block of input data.
- * @param[out] pDst points to the block of output data.
- * @param[in] blockSize number of samples to process.
- */
- void riscv_iir_lattice_f32(
- const riscv_iir_lattice_instance_f32 * S,
- const float32_t * pSrc,
- float32_t * pDst,
- uint32_t blockSize);
-
-
- /**
- * @brief Initialization function for the floating-point IIR lattice filter.
- * @param[in] S points to an instance of the floating-point IIR lattice structure.
- * @param[in] numStages number of stages in the filter.
- * @param[in] pkCoeffs points to the reflection coefficient buffer. The array is of length numStages.
- * @param[in] pvCoeffs points to the ladder coefficient buffer. The array is of length numStages+1.
- * @param[in] pState points to the state buffer. The array is of length numStages+blockSize-1.
- * @param[in] blockSize number of samples to process.
- */
- void riscv_iir_lattice_init_f32(
- riscv_iir_lattice_instance_f32 * S,
- uint16_t numStages,
- float32_t * pkCoeffs,
- float32_t * pvCoeffs,
- float32_t * pState,
- uint32_t blockSize);
-
-
- /**
- * @brief Processing function for the Q31 IIR lattice filter.
- * @param[in] S points to an instance of the Q31 IIR lattice structure.
- * @param[in] pSrc points to the block of input data.
- * @param[out] pDst points to the block of output data.
- * @param[in] blockSize number of samples to process.
- */
- void riscv_iir_lattice_q31(
- const riscv_iir_lattice_instance_q31 * S,
- const q31_t * pSrc,
- q31_t * pDst,
- uint32_t blockSize);
-
-
- /**
- * @brief Initialization function for the Q31 IIR lattice filter.
- * @param[in] S points to an instance of the Q31 IIR lattice structure.
- * @param[in] numStages number of stages in the filter.
- * @param[in] pkCoeffs points to the reflection coefficient buffer. The array is of length numStages.
- * @param[in] pvCoeffs points to the ladder coefficient buffer. The array is of length numStages+1.
- * @param[in] pState points to the state buffer. The array is of length numStages+blockSize.
- * @param[in] blockSize number of samples to process.
- */
- void riscv_iir_lattice_init_q31(
- riscv_iir_lattice_instance_q31 * S,
- uint16_t numStages,
- q31_t * pkCoeffs,
- q31_t * pvCoeffs,
- q31_t * pState,
- uint32_t blockSize);
-
-
- /**
- * @brief Processing function for the Q15 IIR lattice filter.
- * @param[in] S points to an instance of the Q15 IIR lattice structure.
- * @param[in] pSrc points to the block of input data.
- * @param[out] pDst points to the block of output data.
- * @param[in] blockSize number of samples to process.
- */
- void riscv_iir_lattice_q15(
- const riscv_iir_lattice_instance_q15 * S,
- const q15_t * pSrc,
- q15_t * pDst,
- uint32_t blockSize);
-
-
-/**
- * @brief Initialization function for the Q15 IIR lattice filter.
- * @param[in] S points to an instance of the fixed-point Q15 IIR lattice structure.
- * @param[in] numStages number of stages in the filter.
- * @param[in] pkCoeffs points to reflection coefficient buffer. The array is of length numStages.
- * @param[in] pvCoeffs points to ladder coefficient buffer. The array is of length numStages+1.
- * @param[in] pState points to state buffer. The array is of length numStages+blockSize.
- * @param[in] blockSize number of samples to process per call.
- */
- void riscv_iir_lattice_init_q15(
- riscv_iir_lattice_instance_q15 * S,
- uint16_t numStages,
- q15_t * pkCoeffs,
- q15_t * pvCoeffs,
- q15_t * pState,
- uint32_t blockSize);
-
-
- /**
- * @brief Instance structure for the floating-point LMS filter.
- */
- typedef struct
- {
- uint16_t numTaps; /**< number of coefficients in the filter. */
- float32_t *pState; /**< points to the state variable array. The array is of length numTaps+blockSize-1. */
- float32_t *pCoeffs; /**< points to the coefficient array. The array is of length numTaps. */
- float32_t mu; /**< step size that controls filter coefficient updates. */
- } riscv_lms_instance_f32;
-
-
- /**
- * @brief Processing function for floating-point LMS filter.
- * @param[in] S points to an instance of the floating-point LMS filter structure.
- * @param[in] pSrc points to the block of input data.
- * @param[in] pRef points to the block of reference data.
- * @param[out] pOut points to the block of output data.
- * @param[out] pErr points to the block of error data.
- * @param[in] blockSize number of samples to process.
- */
- void riscv_lms_f32(
- const riscv_lms_instance_f32 * S,
- const float32_t * pSrc,
- float32_t * pRef,
- float32_t * pOut,
- float32_t * pErr,
- uint32_t blockSize);
-
-
- /**
- * @brief Initialization function for floating-point LMS filter.
- * @param[in] S points to an instance of the floating-point LMS filter structure.
- * @param[in] numTaps number of filter coefficients.
- * @param[in] pCoeffs points to the coefficient buffer.
- * @param[in] pState points to state buffer.
- * @param[in] mu step size that controls filter coefficient updates.
- * @param[in] blockSize number of samples to process.
- */
- void riscv_lms_init_f32(
- riscv_lms_instance_f32 * S,
- uint16_t numTaps,
- float32_t * pCoeffs,
- float32_t * pState,
- float32_t mu,
- uint32_t blockSize);
-
-
- /**
- * @brief Instance structure for the Q15 LMS filter.
- */
- typedef struct
- {
- uint16_t numTaps; /**< number of coefficients in the filter. */
- q15_t *pState; /**< points to the state variable array. The array is of length numTaps+blockSize-1. */
- q15_t *pCoeffs; /**< points to the coefficient array. The array is of length numTaps. */
- q15_t mu; /**< step size that controls filter coefficient updates. */
- uint32_t postShift; /**< bit shift applied to coefficients. */
- } riscv_lms_instance_q15;
-
-
- /**
- * @brief Initialization function for the Q15 LMS filter.
- * @param[in] S points to an instance of the Q15 LMS filter structure.
- * @param[in] numTaps number of filter coefficients.
- * @param[in] pCoeffs points to the coefficient buffer.
- * @param[in] pState points to the state buffer.
- * @param[in] mu step size that controls filter coefficient updates.
- * @param[in] blockSize number of samples to process.
- * @param[in] postShift bit shift applied to coefficients.
- */
- void riscv_lms_init_q15(
- riscv_lms_instance_q15 * S,
- uint16_t numTaps,
- q15_t * pCoeffs,
- q15_t * pState,
- q15_t mu,
- uint32_t blockSize,
- uint32_t postShift);
-
-
- /**
- * @brief Processing function for Q15 LMS filter.
- * @param[in] S points to an instance of the Q15 LMS filter structure.
- * @param[in] pSrc points to the block of input data.
- * @param[in] pRef points to the block of reference data.
- * @param[out] pOut points to the block of output data.
- * @param[out] pErr points to the block of error data.
- * @param[in] blockSize number of samples to process.
- */
- void riscv_lms_q15(
- const riscv_lms_instance_q15 * S,
- const q15_t * pSrc,
- q15_t * pRef,
- q15_t * pOut,
- q15_t * pErr,
- uint32_t blockSize);
-
-
- /**
- * @brief Instance structure for the Q31 LMS filter.
- */
- typedef struct
- {
- uint16_t numTaps; /**< number of coefficients in the filter. */
- q31_t *pState; /**< points to the state variable array. The array is of length numTaps+blockSize-1. */
- q31_t *pCoeffs; /**< points to the coefficient array. The array is of length numTaps. */
- q31_t mu; /**< step size that controls filter coefficient updates. */
- uint32_t postShift; /**< bit shift applied to coefficients. */
- } riscv_lms_instance_q31;
-
-
- /**
- * @brief Processing function for Q31 LMS filter.
- * @param[in] S points to an instance of the Q15 LMS filter structure.
- * @param[in] pSrc points to the block of input data.
- * @param[in] pRef points to the block of reference data.
- * @param[out] pOut points to the block of output data.
- * @param[out] pErr points to the block of error data.
- * @param[in] blockSize number of samples to process.
- */
- void riscv_lms_q31(
- const riscv_lms_instance_q31 * S,
- const q31_t * pSrc,
- q31_t * pRef,
- q31_t * pOut,
- q31_t * pErr,
- uint32_t blockSize);
-
-
- /**
- * @brief Initialization function for Q31 LMS filter.
- * @param[in] S points to an instance of the Q31 LMS filter structure.
- * @param[in] numTaps number of filter coefficients.
- * @param[in] pCoeffs points to coefficient buffer.
- * @param[in] pState points to state buffer.
- * @param[in] mu step size that controls filter coefficient updates.
- * @param[in] blockSize number of samples to process.
- * @param[in] postShift bit shift applied to coefficients.
- */
- void riscv_lms_init_q31(
- riscv_lms_instance_q31 * S,
- uint16_t numTaps,
- q31_t * pCoeffs,
- q31_t * pState,
- q31_t mu,
- uint32_t blockSize,
- uint32_t postShift);
-
-
- /**
- * @brief Instance structure for the floating-point normalized LMS filter.
- */
- typedef struct
- {
- uint16_t numTaps; /**< number of coefficients in the filter. */
- float32_t *pState; /**< points to the state variable array. The array is of length numTaps+blockSize-1. */
- float32_t *pCoeffs; /**< points to the coefficient array. The array is of length numTaps. */
- float32_t mu; /**< step size that control filter coefficient updates. */
- float32_t energy; /**< saves previous frame energy. */
- float32_t x0; /**< saves previous input sample. */
- } riscv_lms_norm_instance_f32;
-
-
- /**
- * @brief Processing function for floating-point normalized LMS filter.
- * @param[in] S points to an instance of the floating-point normalized LMS filter structure.
- * @param[in] pSrc points to the block of input data.
- * @param[in] pRef points to the block of reference data.
- * @param[out] pOut points to the block of output data.
- * @param[out] pErr points to the block of error data.
- * @param[in] blockSize number of samples to process.
- */
- void riscv_lms_norm_f32(
- riscv_lms_norm_instance_f32 * S,
- const float32_t * pSrc,
- float32_t * pRef,
- float32_t * pOut,
- float32_t * pErr,
- uint32_t blockSize);
-
-
- /**
- * @brief Initialization function for floating-point normalized LMS filter.
- * @param[in] S points to an instance of the floating-point LMS filter structure.
- * @param[in] numTaps number of filter coefficients.
- * @param[in] pCoeffs points to coefficient buffer.
- * @param[in] pState points to state buffer.
- * @param[in] mu step size that controls filter coefficient updates.
- * @param[in] blockSize number of samples to process.
- */
- void riscv_lms_norm_init_f32(
- riscv_lms_norm_instance_f32 * S,
- uint16_t numTaps,
- float32_t * pCoeffs,
- float32_t * pState,
- float32_t mu,
- uint32_t blockSize);
-
-
- /**
- * @brief Instance structure for the Q31 normalized LMS filter.
- */
- typedef struct
- {
- uint16_t numTaps; /**< number of coefficients in the filter. */
- q31_t *pState; /**< points to the state variable array. The array is of length numTaps+blockSize-1. */
- q31_t *pCoeffs; /**< points to the coefficient array. The array is of length numTaps. */
- q31_t mu; /**< step size that controls filter coefficient updates. */
- uint8_t postShift; /**< bit shift applied to coefficients. */
- const q31_t *recipTable; /**< points to the reciprocal initial value table. */
- q31_t energy; /**< saves previous frame energy. */
- q31_t x0; /**< saves previous input sample. */
- } riscv_lms_norm_instance_q31;
-
-
- /**
- * @brief Processing function for Q31 normalized LMS filter.
- * @param[in] S points to an instance of the Q31 normalized LMS filter structure.
- * @param[in] pSrc points to the block of input data.
- * @param[in] pRef points to the block of reference data.
- * @param[out] pOut points to the block of output data.
- * @param[out] pErr points to the block of error data.
- * @param[in] blockSize number of samples to process.
- */
- void riscv_lms_norm_q31(
- riscv_lms_norm_instance_q31 * S,
- const q31_t * pSrc,
- q31_t * pRef,
- q31_t * pOut,
- q31_t * pErr,
- uint32_t blockSize);
-
-
- /**
- * @brief Initialization function for Q31 normalized LMS filter.
- * @param[in] S points to an instance of the Q31 normalized LMS filter structure.
- * @param[in] numTaps number of filter coefficients.
- * @param[in] pCoeffs points to coefficient buffer.
- * @param[in] pState points to state buffer.
- * @param[in] mu step size that controls filter coefficient updates.
- * @param[in] blockSize number of samples to process.
- * @param[in] postShift bit shift applied to coefficients.
- */
- void riscv_lms_norm_init_q31(
- riscv_lms_norm_instance_q31 * S,
- uint16_t numTaps,
- q31_t * pCoeffs,
- q31_t * pState,
- q31_t mu,
- uint32_t blockSize,
- uint8_t postShift);
-
-
- /**
- * @brief Instance structure for the Q15 normalized LMS filter.
- */
- typedef struct
- {
- uint16_t numTaps; /**< Number of coefficients in the filter. */
- q15_t *pState; /**< points to the state variable array. The array is of length numTaps+blockSize-1. */
- q15_t *pCoeffs; /**< points to the coefficient array. The array is of length numTaps. */
- q15_t mu; /**< step size that controls filter coefficient updates. */
- uint8_t postShift; /**< bit shift applied to coefficients. */
- const q15_t *recipTable; /**< Points to the reciprocal initial value table. */
- q15_t energy; /**< saves previous frame energy. */
- q15_t x0; /**< saves previous input sample. */
- } riscv_lms_norm_instance_q15;
-
-
- /**
- * @brief Processing function for Q15 normalized LMS filter.
- * @param[in] S points to an instance of the Q15 normalized LMS filter structure.
- * @param[in] pSrc points to the block of input data.
- * @param[in] pRef points to the block of reference data.
- * @param[out] pOut points to the block of output data.
- * @param[out] pErr points to the block of error data.
- * @param[in] blockSize number of samples to process.
- */
- void riscv_lms_norm_q15(
- riscv_lms_norm_instance_q15 * S,
- const q15_t * pSrc,
- q15_t * pRef,
- q15_t * pOut,
- q15_t * pErr,
- uint32_t blockSize);
-
-
- /**
- * @brief Initialization function for Q15 normalized LMS filter.
- * @param[in] S points to an instance of the Q15 normalized LMS filter structure.
- * @param[in] numTaps number of filter coefficients.
- * @param[in] pCoeffs points to coefficient buffer.
- * @param[in] pState points to state buffer.
- * @param[in] mu step size that controls filter coefficient updates.
- * @param[in] blockSize number of samples to process.
- * @param[in] postShift bit shift applied to coefficients.
- */
- void riscv_lms_norm_init_q15(
- riscv_lms_norm_instance_q15 * S,
- uint16_t numTaps,
- q15_t * pCoeffs,
- q15_t * pState,
- q15_t mu,
- uint32_t blockSize,
- uint8_t postShift);
-
-
- /**
- * @brief Correlation of floating-point sequences.
- * @param[in] pSrcA points to the first input sequence.
- * @param[in] srcALen length of the first input sequence.
- * @param[in] pSrcB points to the second input sequence.
- * @param[in] srcBLen length of the second input sequence.
- * @param[out] pDst points to the block of output data Length 2 * max(srcALen, srcBLen) - 1.
- */
- void riscv_correlate_f32(
- const float32_t * pSrcA,
- uint32_t srcALen,
- const float32_t * pSrcB,
- uint32_t srcBLen,
- float32_t * pDst);
-
-
-/**
- @brief Correlation of Q15 sequences
- @param[in] pSrcA points to the first input sequence
- @param[in] srcALen length of the first input sequence
- @param[in] pSrcB points to the second input sequence
- @param[in] srcBLen length of the second input sequence
- @param[out] pDst points to the block of output data Length 2 * max(srcALen, srcBLen) - 1.
- @param[in] pScratch points to scratch buffer of size max(srcALen, srcBLen) + 2*min(srcALen, srcBLen) - 2.
-*/
-void riscv_correlate_opt_q15(
- const q15_t * pSrcA,
- uint32_t srcALen,
- const q15_t * pSrcB,
- uint32_t srcBLen,
- q15_t * pDst,
- q15_t * pScratch);
-
-
-/**
- @brief Correlation of Q15 sequences.
- @param[in] pSrcA points to the first input sequence
- @param[in] srcALen length of the first input sequence
- @param[in] pSrcB points to the second input sequence
- @param[in] srcBLen length of the second input sequence
- @param[out] pDst points to the block of output data Length 2 * max(srcALen, srcBLen) - 1.
- */
- void riscv_correlate_q15(
- const q15_t * pSrcA,
- uint32_t srcALen,
- const q15_t * pSrcB,
- uint32_t srcBLen,
- q15_t * pDst);
-
-
-/**
- @brief Correlation of Q15 sequences (fast version).
- @param[in] pSrcA points to the first input sequence
- @param[in] srcALen length of the first input sequence
- @param[in] pSrcB points to the second input sequence
- @param[in] srcBLen length of the second input sequence
- @param[out] pDst points to the location where the output result is written. Length 2 * max(srcALen, srcBLen) - 1.
- @return none
- */
-void riscv_correlate_fast_q15(
- const q15_t * pSrcA,
- uint32_t srcALen,
- const q15_t * pSrcB,
- uint32_t srcBLen,
- q15_t * pDst);
-
-/**
- @brief Correlation of Q15 sequences (fast version).
- @param[in] pSrcA points to the first input sequence.
- @param[in] srcALen length of the first input sequence.
- @param[in] pSrcB points to the second input sequence.
- @param[in] srcBLen length of the second input sequence.
- @param[out] pDst points to the block of output data Length 2 * max(srcALen, srcBLen) - 1.
- @param[in] pScratch points to scratch buffer of size max(srcALen, srcBLen) + 2*min(srcALen, srcBLen) - 2.
- */
-void riscv_correlate_fast_opt_q15(
- const q15_t * pSrcA,
- uint32_t srcALen,
- const q15_t * pSrcB,
- uint32_t srcBLen,
- q15_t * pDst,
- q15_t * pScratch);
-
-
- /**
- * @brief Correlation of Q31 sequences.
- * @param[in] pSrcA points to the first input sequence.
- * @param[in] srcALen length of the first input sequence.
- * @param[in] pSrcB points to the second input sequence.
- * @param[in] srcBLen length of the second input sequence.
- * @param[out] pDst points to the block of output data Length 2 * max(srcALen, srcBLen) - 1.
- */
- void riscv_correlate_q31(
- const q31_t * pSrcA,
- uint32_t srcALen,
- const q31_t * pSrcB,
- uint32_t srcBLen,
- q31_t * pDst);
-
-
-/**
- @brief Correlation of Q31 sequences (fast version).
- @param[in] pSrcA points to the first input sequence
- @param[in] srcALen length of the first input sequence
- @param[in] pSrcB points to the second input sequence
- @param[in] srcBLen length of the second input sequence
- @param[out] pDst points to the block of output data Length 2 * max(srcALen, srcBLen) - 1.
- */
-void riscv_correlate_fast_q31(
- const q31_t * pSrcA,
- uint32_t srcALen,
- const q31_t * pSrcB,
- uint32_t srcBLen,
- q31_t * pDst);
-
-
- /**
- * @brief Correlation of Q7 sequences.
- * @param[in] pSrcA points to the first input sequence.
- * @param[in] srcALen length of the first input sequence.
- * @param[in] pSrcB points to the second input sequence.
- * @param[in] srcBLen length of the second input sequence.
- * @param[out] pDst points to the block of output data Length 2 * max(srcALen, srcBLen) - 1.
- * @param[in] pScratch1 points to scratch buffer(of type q15_t) of size max(srcALen, srcBLen) + 2*min(srcALen, srcBLen) - 2.
- * @param[in] pScratch2 points to scratch buffer (of type q15_t) of size min(srcALen, srcBLen).
- */
- void riscv_correlate_opt_q7(
- const q7_t * pSrcA,
- uint32_t srcALen,
- const q7_t * pSrcB,
- uint32_t srcBLen,
- q7_t * pDst,
- q15_t * pScratch1,
- q15_t * pScratch2);
-
-
- /**
- * @brief Correlation of Q7 sequences.
- * @param[in] pSrcA points to the first input sequence.
- * @param[in] srcALen length of the first input sequence.
- * @param[in] pSrcB points to the second input sequence.
- * @param[in] srcBLen length of the second input sequence.
- * @param[out] pDst points to the block of output data Length 2 * max(srcALen, srcBLen) - 1.
- */
- void riscv_correlate_q7(
- const q7_t * pSrcA,
- uint32_t srcALen,
- const q7_t * pSrcB,
- uint32_t srcBLen,
- q7_t * pDst);
-
-
- /**
- * @brief Instance structure for the floating-point sparse FIR filter.
- */
- typedef struct
- {
- uint16_t numTaps; /**< number of coefficients in the filter. */
- uint16_t stateIndex; /**< state buffer index. Points to the oldest sample in the state buffer. */
- float32_t *pState; /**< points to the state buffer array. The array is of length maxDelay+blockSize-1. */
- const float32_t *pCoeffs; /**< points to the coefficient array. The array is of length numTaps.*/
- uint16_t maxDelay; /**< maximum offset specified by the pTapDelay array. */
- int32_t *pTapDelay; /**< points to the array of delay values. The array is of length numTaps. */
- } riscv_fir_sparse_instance_f32;
-
- /**
- * @brief Instance structure for the Q31 sparse FIR filter.
- */
- typedef struct
- {
- uint16_t numTaps; /**< number of coefficients in the filter. */
- uint16_t stateIndex; /**< state buffer index. Points to the oldest sample in the state buffer. */
- q31_t *pState; /**< points to the state buffer array. The array is of length maxDelay+blockSize-1. */
- const q31_t *pCoeffs; /**< points to the coefficient array. The array is of length numTaps.*/
- uint16_t maxDelay; /**< maximum offset specified by the pTapDelay array. */
- int32_t *pTapDelay; /**< points to the array of delay values. The array is of length numTaps. */
- } riscv_fir_sparse_instance_q31;
-
- /**
- * @brief Instance structure for the Q15 sparse FIR filter.
- */
- typedef struct
- {
- uint16_t numTaps; /**< number of coefficients in the filter. */
- uint16_t stateIndex; /**< state buffer index. Points to the oldest sample in the state buffer. */
- q15_t *pState; /**< points to the state buffer array. The array is of length maxDelay+blockSize-1. */
- const q15_t *pCoeffs; /**< points to the coefficient array. The array is of length numTaps.*/
- uint16_t maxDelay; /**< maximum offset specified by the pTapDelay array. */
- int32_t *pTapDelay; /**< points to the array of delay values. The array is of length numTaps. */
- } riscv_fir_sparse_instance_q15;
-
- /**
- * @brief Instance structure for the Q7 sparse FIR filter.
- */
- typedef struct
- {
- uint16_t numTaps; /**< number of coefficients in the filter. */
- uint16_t stateIndex; /**< state buffer index. Points to the oldest sample in the state buffer. */
- q7_t *pState; /**< points to the state buffer array. The array is of length maxDelay+blockSize-1. */
- const q7_t *pCoeffs; /**< points to the coefficient array. The array is of length numTaps.*/
- uint16_t maxDelay; /**< maximum offset specified by the pTapDelay array. */
- int32_t *pTapDelay; /**< points to the array of delay values. The array is of length numTaps. */
- } riscv_fir_sparse_instance_q7;
-
-
- /**
- * @brief Processing function for the floating-point sparse FIR filter.
- * @param[in] S points to an instance of the floating-point sparse FIR structure.
- * @param[in] pSrc points to the block of input data.
- * @param[out] pDst points to the block of output data
- * @param[in] pScratchIn points to a temporary buffer of size blockSize.
- * @param[in] blockSize number of input samples to process per call.
- */
- void riscv_fir_sparse_f32(
- riscv_fir_sparse_instance_f32 * S,
- const float32_t * pSrc,
- float32_t * pDst,
- float32_t * pScratchIn,
- uint32_t blockSize);
-
-
- /**
- * @brief Initialization function for the floating-point sparse FIR filter.
- * @param[in,out] S points to an instance of the floating-point sparse FIR structure.
- * @param[in] numTaps number of nonzero coefficients in the filter.
- * @param[in] pCoeffs points to the array of filter coefficients.
- * @param[in] pState points to the state buffer.
- * @param[in] pTapDelay points to the array of offset times.
- * @param[in] maxDelay maximum offset time supported.
- * @param[in] blockSize number of samples that will be processed per block.
- */
- void riscv_fir_sparse_init_f32(
- riscv_fir_sparse_instance_f32 * S,
- uint16_t numTaps,
- const float32_t * pCoeffs,
- float32_t * pState,
- int32_t * pTapDelay,
- uint16_t maxDelay,
- uint32_t blockSize);
-
-
- /**
- * @brief Processing function for the Q31 sparse FIR filter.
- * @param[in] S points to an instance of the Q31 sparse FIR structure.
- * @param[in] pSrc points to the block of input data.
- * @param[out] pDst points to the block of output data
- * @param[in] pScratchIn points to a temporary buffer of size blockSize.
- * @param[in] blockSize number of input samples to process per call.
- */
- void riscv_fir_sparse_q31(
- riscv_fir_sparse_instance_q31 * S,
- const q31_t * pSrc,
- q31_t * pDst,
- q31_t * pScratchIn,
- uint32_t blockSize);
-
-
- /**
- * @brief Initialization function for the Q31 sparse FIR filter.
- * @param[in,out] S points to an instance of the Q31 sparse FIR structure.
- * @param[in] numTaps number of nonzero coefficients in the filter.
- * @param[in] pCoeffs points to the array of filter coefficients.
- * @param[in] pState points to the state buffer.
- * @param[in] pTapDelay points to the array of offset times.
- * @param[in] maxDelay maximum offset time supported.
- * @param[in] blockSize number of samples that will be processed per block.
- */
- void riscv_fir_sparse_init_q31(
- riscv_fir_sparse_instance_q31 * S,
- uint16_t numTaps,
- const q31_t * pCoeffs,
- q31_t * pState,
- int32_t * pTapDelay,
- uint16_t maxDelay,
- uint32_t blockSize);
-
-
- /**
- * @brief Processing function for the Q15 sparse FIR filter.
- * @param[in] S points to an instance of the Q15 sparse FIR structure.
- * @param[in] pSrc points to the block of input data.
- * @param[out] pDst points to the block of output data
- * @param[in] pScratchIn points to a temporary buffer of size blockSize.
- * @param[in] pScratchOut points to a temporary buffer of size blockSize.
- * @param[in] blockSize number of input samples to process per call.
- */
- void riscv_fir_sparse_q15(
- riscv_fir_sparse_instance_q15 * S,
- const q15_t * pSrc,
- q15_t * pDst,
- q15_t * pScratchIn,
- q31_t * pScratchOut,
- uint32_t blockSize);
-
-
- /**
- * @brief Initialization function for the Q15 sparse FIR filter.
- * @param[in,out] S points to an instance of the Q15 sparse FIR structure.
- * @param[in] numTaps number of nonzero coefficients in the filter.
- * @param[in] pCoeffs points to the array of filter coefficients.
- * @param[in] pState points to the state buffer.
- * @param[in] pTapDelay points to the array of offset times.
- * @param[in] maxDelay maximum offset time supported.
- * @param[in] blockSize number of samples that will be processed per block.
- */
- void riscv_fir_sparse_init_q15(
- riscv_fir_sparse_instance_q15 * S,
- uint16_t numTaps,
- const q15_t * pCoeffs,
- q15_t * pState,
- int32_t * pTapDelay,
- uint16_t maxDelay,
- uint32_t blockSize);
-
-
- /**
- * @brief Processing function for the Q7 sparse FIR filter.
- * @param[in] S points to an instance of the Q7 sparse FIR structure.
- * @param[in] pSrc points to the block of input data.
- * @param[out] pDst points to the block of output data
- * @param[in] pScratchIn points to a temporary buffer of size blockSize.
- * @param[in] pScratchOut points to a temporary buffer of size blockSize.
- * @param[in] blockSize number of input samples to process per call.
- */
- void riscv_fir_sparse_q7(
- riscv_fir_sparse_instance_q7 * S,
- const q7_t * pSrc,
- q7_t * pDst,
- q7_t * pScratchIn,
- q31_t * pScratchOut,
- uint32_t blockSize);
-
-
- /**
- * @brief Initialization function for the Q7 sparse FIR filter.
- * @param[in,out] S points to an instance of the Q7 sparse FIR structure.
- * @param[in] numTaps number of nonzero coefficients in the filter.
- * @param[in] pCoeffs points to the array of filter coefficients.
- * @param[in] pState points to the state buffer.
- * @param[in] pTapDelay points to the array of offset times.
- * @param[in] maxDelay maximum offset time supported.
- * @param[in] blockSize number of samples that will be processed per block.
- */
- void riscv_fir_sparse_init_q7(
- riscv_fir_sparse_instance_q7 * S,
- uint16_t numTaps,
- const q7_t * pCoeffs,
- q7_t * pState,
- int32_t * pTapDelay,
- uint16_t maxDelay,
- uint32_t blockSize);
-
-
- /**
- * @brief Floating-point sin_cos function.
- * @param[in] theta input value in degrees
- * @param[out] pSinVal points to the processed sine output.
- * @param[out] pCosVal points to the processed cos output.
- */
- void riscv_sin_cos_f32(
- float32_t theta,
- float32_t * pSinVal,
- float32_t * pCosVal);
-
-
- /**
- * @brief Q31 sin_cos function.
- * @param[in] theta scaled input value in degrees
- * @param[out] pSinVal points to the processed sine output.
- * @param[out] pCosVal points to the processed cosine output.
- */
- void riscv_sin_cos_q31(
- q31_t theta,
- q31_t * pSinVal,
- q31_t * pCosVal);
-
-
- /**
- * @brief Floating-point complex conjugate.
- * @param[in] pSrc points to the input vector
- * @param[out] pDst points to the output vector
- * @param[in] numSamples number of complex samples in each vector
- */
- void riscv_cmplx_conj_f32(
- const float32_t * pSrc,
- float32_t * pDst,
- uint32_t numSamples);
-
- /**
- * @brief Q31 complex conjugate.
- * @param[in] pSrc points to the input vector
- * @param[out] pDst points to the output vector
- * @param[in] numSamples number of complex samples in each vector
- */
- void riscv_cmplx_conj_q31(
- const q31_t * pSrc,
- q31_t * pDst,
- uint32_t numSamples);
-
-
- /**
- * @brief Q15 complex conjugate.
- * @param[in] pSrc points to the input vector
- * @param[out] pDst points to the output vector
- * @param[in] numSamples number of complex samples in each vector
- */
- void riscv_cmplx_conj_q15(
- const q15_t * pSrc,
- q15_t * pDst,
- uint32_t numSamples);
-
-
- /**
- * @brief Floating-point complex magnitude squared
- * @param[in] pSrc points to the complex input vector
- * @param[out] pDst points to the real output vector
- * @param[in] numSamples number of complex samples in the input vector
- */
- void riscv_cmplx_mag_squared_f32(
- const float32_t * pSrc,
- float32_t * pDst,
- uint32_t numSamples);
-
-
- /**
- * @brief Q31 complex magnitude squared
- * @param[in] pSrc points to the complex input vector
- * @param[out] pDst points to the real output vector
- * @param[in] numSamples number of complex samples in the input vector
- */
- void riscv_cmplx_mag_squared_q31(
- const q31_t * pSrc,
- q31_t * pDst,
- uint32_t numSamples);
-
-
- /**
- * @brief Q15 complex magnitude squared
- * @param[in] pSrc points to the complex input vector
- * @param[out] pDst points to the real output vector
- * @param[in] numSamples number of complex samples in the input vector
- */
- void riscv_cmplx_mag_squared_q15(
- const q15_t * pSrc,
- q15_t * pDst,
- uint32_t numSamples);
-
-
- /**
- * @ingroup groupController
- */
-
- /**
- * @defgroup PID PID Motor Control
- *
- * A Proportional Integral Derivative (PID) controller is a generic feedback control
- * loop mechanism widely used in industrial control systems.
- * A PID controller is the most commonly used type of feedback controller.
- *
- * This set of functions implements (PID) controllers
- * for Q15, Q31, and floating-point data types. The functions operate on a single sample
- * of data and each call to the function returns a single processed value.
- * S points to an instance of the PID control data structure. in
- * is the input sample value. The functions return the output value.
- *
- * \par Algorithm:
- *
- * y[n] = y[n-1] + A0 * x[n] + A1 * x[n-1] + A2 * x[n-2]
- * A0 = Kp + Ki + Kd
- * A1 = (-Kp ) - (2 * Kd )
- * A2 = Kd
- *
- *
- * \par
- * where \c Kp is proportional constant, \c Ki is Integral constant and \c Kd is Derivative constant
- *
- * \par
- * \image html PID.png "Proportional Integral Derivative Controller"
- *
- * \par
- * The PID controller calculates an "error" value as the difference between
- * the measured output and the reference input.
- * The controller attempts to minimize the error by adjusting the process control inputs.
- * The proportional value determines the reaction to the current error,
- * the integral value determines the reaction based on the sum of recent errors,
- * and the derivative value determines the reaction based on the rate at which the error has been changing.
- *
- * \par Instance Structure
- * The Gains A0, A1, A2 and state variables for a PID controller are stored together in an instance data structure.
- * A separate instance structure must be defined for each PID Controller.
- * There are separate instance structure declarations for each of the 3 supported data types.
- *
- * \par Reset Functions
- * There is also an associated reset function for each data type which clears the state array.
- *
- * \par Initialization Functions
- * There is also an associated initialization function for each data type.
- * The initialization function performs the following operations:
- * - Initializes the Gains A0, A1, A2 from Kp,Ki, Kd gains.
- * - Zeros out the values in the state buffer.
- *
- * \par
- * Instance structure cannot be placed into a const data section and it is recommended to use the initialization function.
- *
- * \par Fixed-Point Behavior
- * Care must be taken when using the fixed-point versions of the PID Controller functions.
- * In particular, the overflow and saturation behavior of the accumulator used in each function must be considered.
- * Refer to the function specific documentation below for usage guidelines.
- */
-
- /**
- * @addtogroup PID
- * @{
- */
-
- /**
- * @brief Process function for the floating-point PID Control.
- * @param[in,out] S is an instance of the floating-point PID Control structure
- * @param[in] in input sample to process
- * @return processed output sample.
- */
- __STATIC_FORCEINLINE float32_t riscv_pid_f32(
- riscv_pid_instance_f32 * S,
- float32_t in)
- {
- float32_t out;
-
- /* y[n] = y[n-1] + A0 * x[n] + A1 * x[n-1] + A2 * x[n-2] */
- out = (S->A0 * in) +
- (S->A1 * S->state[0]) + (S->A2 * S->state[1]) + (S->state[2]);
-
- /* Update state */
- S->state[1] = S->state[0];
- S->state[0] = in;
- S->state[2] = out;
-
- /* return to application */
- return (out);
-
- }
-
-/**
- @brief Process function for the Q31 PID Control.
- @param[in,out] S points to an instance of the Q31 PID Control structure
- @param[in] in input sample to process
- @return processed output sample.
-
- \par Scaling and Overflow Behavior
- The function is implemented using an internal 64-bit accumulator.
- The accumulator has a 2.62 format and maintains full precision of the intermediate multiplication results but provides only a single guard bit.
- Thus, if the accumulator result overflows it wraps around rather than clip.
- In order to avoid overflows completely the input signal must be scaled down by 2 bits as there are four additions.
- After all multiply-accumulates are performed, the 2.62 accumulator is truncated to 1.32 format and then saturated to 1.31 format.
- */
-__STATIC_FORCEINLINE q31_t riscv_pid_q31(
- riscv_pid_instance_q31 * S,
- q31_t in)
- {
- q63_t acc;
- q31_t out;
-
- /* acc = A0 * x[n] */
- acc = (q63_t) S->A0 * in;
-
- /* acc += A1 * x[n-1] */
- acc += (q63_t) S->A1 * S->state[0];
-
- /* acc += A2 * x[n-2] */
- acc += (q63_t) S->A2 * S->state[1];
-
- /* convert output to 1.31 format to add y[n-1] */
- out = (q31_t) (acc >> 31U);
-
- /* out += y[n-1] */
- out += S->state[2];
-
- /* Update state */
- S->state[1] = S->state[0];
- S->state[0] = in;
- S->state[2] = out;
-
- /* return to application */
- return (out);
- }
-
-
-/**
- @brief Process function for the Q15 PID Control.
- @param[in,out] S points to an instance of the Q15 PID Control structure
- @param[in] in input sample to process
- @return processed output sample.
-
- \par Scaling and Overflow Behavior
- The function is implemented using a 64-bit internal accumulator.
- Both Gains and state variables are represented in 1.15 format and multiplications yield a 2.30 result.
- The 2.30 intermediate results are accumulated in a 64-bit accumulator in 34.30 format.
- There is no risk of internal overflow with this approach and the full precision of intermediate multiplications is preserved.
- After all additions have been performed, the accumulator is truncated to 34.15 format by discarding low 15 bits.
- Lastly, the accumulator is saturated to yield a result in 1.15 format.
- */
-__STATIC_FORCEINLINE q15_t riscv_pid_q15(
- riscv_pid_instance_q15 * S,
- q15_t in)
- {
- q63_t acc;
- q15_t out;
-
-#if defined (RISCV_MATH_DSP)
- /* Implementation of PID controller */
-
- /* acc = A0 * x[n] */
- acc = (q31_t) __RV_KMDA((uint32_t)S->A0, (uint32_t)in);
-
- /* acc += A1 * x[n-1] + A2 * x[n-2] */
- acc = (q63_t)__RV_SMALDA((uint64_t)acc, (uint32_t)S->A1, (uint32_t)read_q15x2 (S->state));
-#else
- /* acc = A0 * x[n] */
- acc = ((q31_t) S->A0) * in;
-
- /* acc += A1 * x[n-1] + A2 * x[n-2] */
- acc += (q31_t) S->A1 * S->state[0];
- acc += (q31_t) S->A2 * S->state[1];
-#endif
-
- /* acc += y[n-1] */
- acc += (q31_t) S->state[2] << 15;
-
- /* saturate the output */
- out = (q15_t) (__SSAT((acc >> 15), 16));
-
- /* Update state */
- S->state[1] = S->state[0];
- S->state[0] = in;
- S->state[2] = out;
-
- /* return to application */
- return (out);
- }
-
- /**
- * @} end of PID group
- */
-
-
- /**
- * @brief Floating-point matrix inverse.
- * @param[in] src points to the instance of the input floating-point matrix structure.
- * @param[out] dst points to the instance of the output floating-point matrix structure.
- * @return The function returns RISCV_MATH_SIZE_MISMATCH, if the dimensions do not match.
- * If the input matrix is singular (does not have an inverse), then the algorithm terminates and returns error status RISCV_MATH_SINGULAR.
- */
- riscv_status riscv_mat_inverse_f32(
- const riscv_matrix_instance_f32 * src,
- riscv_matrix_instance_f32 * dst);
-
-
- /**
- * @brief Floating-point matrix inverse.
- * @param[in] src points to the instance of the input floating-point matrix structure.
- * @param[out] dst points to the instance of the output floating-point matrix structure.
- * @return The function returns RISCV_MATH_SIZE_MISMATCH, if the dimensions do not match.
- * If the input matrix is singular (does not have an inverse), then the algorithm terminates and returns error status RISCV_MATH_SINGULAR.
- */
- riscv_status riscv_mat_inverse_f64(
- const riscv_matrix_instance_f64 * src,
- riscv_matrix_instance_f64 * dst);
-
-
-
- /**
- * @ingroup groupController
- */
-
- /**
- * @defgroup clarke Vector Clarke Transform
- * Forward Clarke transform converts the instantaneous stator phases into a two-coordinate time invariant vector.
- * Generally the Clarke transform uses three-phase currents Ia, Ib and Ic to calculate currents
- * in the two-phase orthogonal stator axis Ialpha and Ibeta.
- * When Ialpha is superposed with Ia as shown in the figure below
- * \image html clarke.png Stator current space vector and its components in (a,b).
- * and Ia + Ib + Ic = 0, in this condition Ialpha and Ibeta
- * can be calculated using only Ia and Ib.
- *
- * The function operates on a single sample of data and each call to the function returns the processed output.
- * The library provides separate functions for Q31 and floating-point data types.
- * \par Algorithm
- * \image html clarkeFormula.png
- * where Ia and Ib are the instantaneous stator phases and
- * pIalpha and pIbeta are the two coordinates of time invariant vector.
- * \par Fixed-Point Behavior
- * Care must be taken when using the Q31 version of the Clarke transform.
- * In particular, the overflow and saturation behavior of the accumulator used must be considered.
- * Refer to the function specific documentation below for usage guidelines.
- */
-
- /**
- * @addtogroup clarke
- * @{
- */
-
- /**
- *
- * @brief Floating-point Clarke transform
- * @param[in] Ia input three-phase coordinate a
- * @param[in] Ib input three-phase coordinate b
- * @param[out] pIalpha points to output two-phase orthogonal vector axis alpha
- * @param[out] pIbeta points to output two-phase orthogonal vector axis beta
- * @return none
- */
- __STATIC_FORCEINLINE void riscv_clarke_f32(
- float32_t Ia,
- float32_t Ib,
- float32_t * pIalpha,
- float32_t * pIbeta)
- {
- /* Calculate pIalpha using the equation, pIalpha = Ia */
- *pIalpha = Ia;
-
- /* Calculate pIbeta using the equation, pIbeta = (1/sqrt(3)) * Ia + (2/sqrt(3)) * Ib */
- *pIbeta = ((float32_t) 0.57735026919 * Ia + (float32_t) 1.15470053838 * Ib);
- }
-
-
-/**
- @brief Clarke transform for Q31 version
- @param[in] Ia input three-phase coordinate a
- @param[in] Ib input three-phase coordinate b
- @param[out] pIalpha points to output two-phase orthogonal vector axis alpha
- @param[out] pIbeta points to output two-phase orthogonal vector axis beta
- @return none
-
- \par Scaling and Overflow Behavior
- The function is implemented using an internal 32-bit accumulator.
- The accumulator maintains 1.31 format by truncating lower 31 bits of the intermediate multiplication in 2.62 format.
- There is saturation on the addition, hence there is no risk of overflow.
- */
-__STATIC_FORCEINLINE void riscv_clarke_q31(
- q31_t Ia,
- q31_t Ib,
- q31_t * pIalpha,
- q31_t * pIbeta)
- {
- q31_t product1, product2; /* Temporary variables used to store intermediate results */
-
- /* Calculating pIalpha from Ia by equation pIalpha = Ia */
- *pIalpha = Ia;
-
- /* Intermediate product is calculated by (1/(sqrt(3)) * Ia) */
- product1 = (q31_t) (((q63_t) Ia * 0x24F34E8B) >> 30);
-
- /* Intermediate product is calculated by (2/sqrt(3) * Ib) */
- product2 = (q31_t) (((q63_t) Ib * 0x49E69D16) >> 30);
-
- /* pIbeta is calculated by adding the intermediate products */
- *pIbeta = __QADD(product1, product2);
- }
-
- /**
- * @} end of clarke group
- */
-
-
- /**
- * @ingroup groupController
- */
-
- /**
- * @defgroup inv_clarke Vector Inverse Clarke Transform
- * Inverse Clarke transform converts the two-coordinate time invariant vector into instantaneous stator phases.
- *
- * The function operates on a single sample of data and each call to the function returns the processed output.
- * The library provides separate functions for Q31 and floating-point data types.
- * \par Algorithm
- * \image html clarkeInvFormula.png
- * where pIa and pIb are the instantaneous stator phases and
- * Ialpha and Ibeta are the two coordinates of time invariant vector.
- * \par Fixed-Point Behavior
- * Care must be taken when using the Q31 version of the Clarke transform.
- * In particular, the overflow and saturation behavior of the accumulator used must be considered.
- * Refer to the function specific documentation below for usage guidelines.
- */
-
- /**
- * @addtogroup inv_clarke
- * @{
- */
-
- /**
- * @brief Floating-point Inverse Clarke transform
- * @param[in] Ialpha input two-phase orthogonal vector axis alpha
- * @param[in] Ibeta input two-phase orthogonal vector axis beta
- * @param[out] pIa points to output three-phase coordinate a
- * @param[out] pIb points to output three-phase coordinate b
- * @return none
- */
- __STATIC_FORCEINLINE void riscv_inv_clarke_f32(
- float32_t Ialpha,
- float32_t Ibeta,
- float32_t * pIa,
- float32_t * pIb)
- {
- /* Calculating pIa from Ialpha by equation pIa = Ialpha */
- *pIa = Ialpha;
-
- /* Calculating pIb from Ialpha and Ibeta by equation pIb = -(1/2) * Ialpha + (sqrt(3)/2) * Ibeta */
- *pIb = -0.5f * Ialpha + 0.8660254039f * Ibeta;
- }
-
-
-/**
- @brief Inverse Clarke transform for Q31 version
- @param[in] Ialpha input two-phase orthogonal vector axis alpha
- @param[in] Ibeta input two-phase orthogonal vector axis beta
- @param[out] pIa points to output three-phase coordinate a
- @param[out] pIb points to output three-phase coordinate b
- @return none
-
- \par Scaling and Overflow Behavior
- The function is implemented using an internal 32-bit accumulator.
- The accumulator maintains 1.31 format by truncating lower 31 bits of the intermediate multiplication in 2.62 format.
- There is saturation on the subtraction, hence there is no risk of overflow.
- */
-__STATIC_FORCEINLINE void riscv_inv_clarke_q31(
- q31_t Ialpha,
- q31_t Ibeta,
- q31_t * pIa,
- q31_t * pIb)
- {
- q31_t product1, product2; /* Temporary variables used to store intermediate results */
-
- /* Calculating pIa from Ialpha by equation pIa = Ialpha */
- *pIa = Ialpha;
-
- /* Intermediate product is calculated by (1/(2*sqrt(3)) * Ia) */
- product1 = (q31_t) (((q63_t) (Ialpha) * (0x40000000)) >> 31);
-
- /* Intermediate product is calculated by (1/sqrt(3) * pIb) */
- product2 = (q31_t) (((q63_t) (Ibeta) * (0x6ED9EBA1)) >> 31);
-
- /* pIb is calculated by subtracting the products */
- *pIb = __QSUB(product2, product1);
- }
-
- /**
- * @} end of inv_clarke group
- */
-
-
-
- /**
- * @ingroup groupController
- */
-
- /**
- * @defgroup park Vector Park Transform
- *
- * Forward Park transform converts the input two-coordinate vector to flux and torque components.
- * The Park transform can be used to realize the transformation of the Ialpha and the Ibeta currents
- * from the stationary to the moving reference frame and control the spatial relationship between
- * the stator vector current and rotor flux vector.
- * If we consider the d axis aligned with the rotor flux, the diagram below shows the
- * current vector and the relationship from the two reference frames:
- * \image html park.png "Stator current space vector and its component in (a,b) and in the d,q rotating reference frame"
- *
- * The function operates on a single sample of data and each call to the function returns the processed output.
- * The library provides separate functions for Q31 and floating-point data types.
- * \par Algorithm
- * \image html parkFormula.png
- * where Ialpha and Ibeta are the stator vector components,
- * pId and pIq are rotor vector components and cosVal and sinVal are the
- * cosine and sine values of theta (rotor flux position).
- * \par Fixed-Point Behavior
- * Care must be taken when using the Q31 version of the Park transform.
- * In particular, the overflow and saturation behavior of the accumulator used must be considered.
- * Refer to the function specific documentation below for usage guidelines.
- */
-
- /**
- * @addtogroup park
- * @{
- */
-
- /**
- * @brief Floating-point Park transform
- * @param[in] Ialpha input two-phase vector coordinate alpha
- * @param[in] Ibeta input two-phase vector coordinate beta
- * @param[out] pId points to output rotor reference frame d
- * @param[out] pIq points to output rotor reference frame q
- * @param[in] sinVal sine value of rotation angle theta
- * @param[in] cosVal cosine value of rotation angle theta
- * @return none
- *
- * The function implements the forward Park transform.
- *
- */
- __STATIC_FORCEINLINE void riscv_park_f32(
- float32_t Ialpha,
- float32_t Ibeta,
- float32_t * pId,
- float32_t * pIq,
- float32_t sinVal,
- float32_t cosVal)
- {
- /* Calculate pId using the equation, pId = Ialpha * cosVal + Ibeta * sinVal */
- *pId = Ialpha * cosVal + Ibeta * sinVal;
-
- /* Calculate pIq using the equation, pIq = - Ialpha * sinVal + Ibeta * cosVal */
- *pIq = -Ialpha * sinVal + Ibeta * cosVal;
- }
-
-
-/**
- @brief Park transform for Q31 version
- @param[in] Ialpha input two-phase vector coordinate alpha
- @param[in] Ibeta input two-phase vector coordinate beta
- @param[out] pId points to output rotor reference frame d
- @param[out] pIq points to output rotor reference frame q
- @param[in] sinVal sine value of rotation angle theta
- @param[in] cosVal cosine value of rotation angle theta
- @return none
-
- \par Scaling and Overflow Behavior
- The function is implemented using an internal 32-bit accumulator.
- The accumulator maintains 1.31 format by truncating lower 31 bits of the intermediate multiplication in 2.62 format.
- There is saturation on the addition and subtraction, hence there is no risk of overflow.
- */
-__STATIC_FORCEINLINE void riscv_park_q31(
- q31_t Ialpha,
- q31_t Ibeta,
- q31_t * pId,
- q31_t * pIq,
- q31_t sinVal,
- q31_t cosVal)
- {
- q31_t product1, product2; /* Temporary variables used to store intermediate results */
- q31_t product3, product4; /* Temporary variables used to store intermediate results */
-
- /* Intermediate product is calculated by (Ialpha * cosVal) */
- product1 = (q31_t) (((q63_t) (Ialpha) * (cosVal)) >> 31);
-
- /* Intermediate product is calculated by (Ibeta * sinVal) */
- product2 = (q31_t) (((q63_t) (Ibeta) * (sinVal)) >> 31);
-
-
- /* Intermediate product is calculated by (Ialpha * sinVal) */
- product3 = (q31_t) (((q63_t) (Ialpha) * (sinVal)) >> 31);
-
- /* Intermediate product is calculated by (Ibeta * cosVal) */
- product4 = (q31_t) (((q63_t) (Ibeta) * (cosVal)) >> 31);
-
- /* Calculate pId by adding the two intermediate products 1 and 2 */
- *pId = __QADD(product1, product2);
-
- /* Calculate pIq by subtracting the two intermediate products 3 from 4 */
- *pIq = __QSUB(product4, product3);
- }
-
- /**
- * @} end of park group
- */
-
-
- /**
- * @ingroup groupController
- */
-
- /**
- * @defgroup inv_park Vector Inverse Park transform
- * Inverse Park transform converts the input flux and torque components to two-coordinate vector.
- *
- * The function operates on a single sample of data and each call to the function returns the processed output.
- * The library provides separate functions for Q31 and floating-point data types.
- * \par Algorithm
- * \image html parkInvFormula.png
- * where pIalpha and pIbeta are the stator vector components,
- * Id and Iq are rotor vector components and cosVal and sinVal are the
- * cosine and sine values of theta (rotor flux position).
- * \par Fixed-Point Behavior
- * Care must be taken when using the Q31 version of the Park transform.
- * In particular, the overflow and saturation behavior of the accumulator used must be considered.
- * Refer to the function specific documentation below for usage guidelines.
- */
-
- /**
- * @addtogroup inv_park
- * @{
- */
-
- /**
- * @brief Floating-point Inverse Park transform
- * @param[in] Id input coordinate of rotor reference frame d
- * @param[in] Iq input coordinate of rotor reference frame q
- * @param[out] pIalpha points to output two-phase orthogonal vector axis alpha
- * @param[out] pIbeta points to output two-phase orthogonal vector axis beta
- * @param[in] sinVal sine value of rotation angle theta
- * @param[in] cosVal cosine value of rotation angle theta
- * @return none
- */
- __STATIC_FORCEINLINE void riscv_inv_park_f32(
- float32_t Id,
- float32_t Iq,
- float32_t * pIalpha,
- float32_t * pIbeta,
- float32_t sinVal,
- float32_t cosVal)
- {
- /* Calculate pIalpha using the equation, pIalpha = Id * cosVal - Iq * sinVal */
- *pIalpha = Id * cosVal - Iq * sinVal;
-
- /* Calculate pIbeta using the equation, pIbeta = Id * sinVal + Iq * cosVal */
- *pIbeta = Id * sinVal + Iq * cosVal;
- }
-
-
-/**
- @brief Inverse Park transform for Q31 version
- @param[in] Id input coordinate of rotor reference frame d
- @param[in] Iq input coordinate of rotor reference frame q
- @param[out] pIalpha points to output two-phase orthogonal vector axis alpha
- @param[out] pIbeta points to output two-phase orthogonal vector axis beta
- @param[in] sinVal sine value of rotation angle theta
- @param[in] cosVal cosine value of rotation angle theta
- @return none
-
- @par Scaling and Overflow Behavior
- The function is implemented using an internal 32-bit accumulator.
- The accumulator maintains 1.31 format by truncating lower 31 bits of the intermediate multiplication in 2.62 format.
- There is saturation on the addition, hence there is no risk of overflow.
- */
-__STATIC_FORCEINLINE void riscv_inv_park_q31(
- q31_t Id,
- q31_t Iq,
- q31_t * pIalpha,
- q31_t * pIbeta,
- q31_t sinVal,
- q31_t cosVal)
- {
- q31_t product1, product2; /* Temporary variables used to store intermediate results */
- q31_t product3, product4; /* Temporary variables used to store intermediate results */
-
- /* Intermediate product is calculated by (Id * cosVal) */
- product1 = (q31_t) (((q63_t) (Id) * (cosVal)) >> 31);
-
- /* Intermediate product is calculated by (Iq * sinVal) */
- product2 = (q31_t) (((q63_t) (Iq) * (sinVal)) >> 31);
-
-
- /* Intermediate product is calculated by (Id * sinVal) */
- product3 = (q31_t) (((q63_t) (Id) * (sinVal)) >> 31);
-
- /* Intermediate product is calculated by (Iq * cosVal) */
- product4 = (q31_t) (((q63_t) (Iq) * (cosVal)) >> 31);
-
- /* Calculate pIalpha by using the two intermediate products 1 and 2 */
- *pIalpha = __QSUB(product1, product2);
-
- /* Calculate pIbeta by using the two intermediate products 3 and 4 */
- *pIbeta = __QADD(product4, product3);
- }
-
- /**
- * @} end of Inverse park group
- */
-
-
- /**
- * @ingroup groupInterpolation
- */
-
- /**
- * @defgroup LinearInterpolate Linear Interpolation
- *
- * Linear interpolation is a method of curve fitting using linear polynomials.
- * Linear interpolation works by effectively drawing a straight line between two neighboring samples and returning the appropriate point along that line
- *
- * \par
- * \image html LinearInterp.png "Linear interpolation"
- *
- * \par
- * A Linear Interpolate function calculates an output value(y), for the input(x)
- * using linear interpolation of the input values x0, x1( nearest input values) and the output values y0 and y1(nearest output values)
- *
- * \par Algorithm:
- *
- * y = y0 + (x - x0) * ((y1 - y0)/(x1-x0))
- * where x0, x1 are nearest values of input x
- * y0, y1 are nearest values to output y
- *
- *
- * \par
- * This set of functions implements Linear interpolation process
- * for Q7, Q15, Q31, and floating-point data types. The functions operate on a single
- * sample of data and each call to the function returns a single processed value.
- * S points to an instance of the Linear Interpolate function data structure.
- * x is the input sample value. The functions returns the output value.
- *
- * \par
- * if x is outside of the table boundary, Linear interpolation returns first value of the table
- * if x is below input range and returns last value of table if x is above range.
- */
-
- /**
- * @addtogroup LinearInterpolate
- * @{
- */
-
- /**
- * @brief Process function for the floating-point Linear Interpolation Function.
- * @param[in,out] S is an instance of the floating-point Linear Interpolation structure
- * @param[in] x input sample to process
- * @return y processed output sample.
- *
- */
- __STATIC_FORCEINLINE float32_t riscv_linear_interp_f32(
- riscv_linear_interp_instance_f32 * S,
- float32_t x)
- {
- float32_t y;
- float32_t x0, x1; /* Nearest input values */
- float32_t y0, y1; /* Nearest output values */
- float32_t xSpacing = S->xSpacing; /* spacing between input values */
- int32_t i; /* Index variable */
- float32_t *pYData = S->pYData; /* pointer to output table */
-
- /* Calculation of index */
- i = (int32_t) ((x - S->x1) / xSpacing);
-
- if (i < 0)
- {
- /* Iniatilize output for below specified range as least output value of table */
- y = pYData[0];
- }
- else if ((uint32_t)i >= (S->nValues - 1))
- {
- /* Iniatilize output for above specified range as last output value of table */
- y = pYData[S->nValues - 1];
- }
- else
- {
- /* Calculation of nearest input values */
- x0 = S->x1 + i * xSpacing;
- x1 = S->x1 + (i + 1) * xSpacing;
-
- /* Read of nearest output values */
- y0 = pYData[i];
- y1 = pYData[i + 1];
-
- /* Calculation of output */
- y = y0 + (x - x0) * ((y1 - y0) / (x1 - x0));
-
- }
-
- /* returns output value */
- return (y);
- }
-
-
- /**
- *
- * @brief Process function for the Q31 Linear Interpolation Function.
- * @param[in] pYData pointer to Q31 Linear Interpolation table
- * @param[in] x input sample to process
- * @param[in] nValues number of table values
- * @return y processed output sample.
- *
- * \par
- * Input sample x is in 12.20 format which contains 12 bits for table index and 20 bits for fractional part.
- * This function can support maximum of table size 2^12.
- *
- */
- __STATIC_FORCEINLINE q31_t riscv_linear_interp_q31(
- q31_t * pYData,
- q31_t x,
- uint32_t nValues)
- {
- q31_t y; /* output */
- q31_t y0, y1; /* Nearest output values */
- q31_t fract; /* fractional part */
- int32_t index; /* Index to read nearest output values */
-
- /* Input is in 12.20 format */
- /* 12 bits for the table index */
- /* Index value calculation */
- index = ((x & (q31_t)0xFFF00000) >> 20);
-
- if (index >= (int32_t)(nValues - 1))
- {
- return (pYData[nValues - 1]);
- }
- else if (index < 0)
- {
- return (pYData[0]);
- }
- else
- {
- /* 20 bits for the fractional part */
- /* shift left by 11 to keep fract in 1.31 format */
- fract = (x & 0x000FFFFF) << 11;
-
- /* Read two nearest output values from the index in 1.31(q31) format */
- y0 = pYData[index];
- y1 = pYData[index + 1];
-
- /* Calculation of y0 * (1-fract) and y is in 2.30 format */
- y = ((q31_t) ((q63_t) y0 * (0x7FFFFFFF - fract) >> 32));
-
- /* Calculation of y0 * (1-fract) + y1 *fract and y is in 2.30 format */
- y += ((q31_t) (((q63_t) y1 * fract) >> 32));
-
- /* Convert y to 1.31 format */
- return (y << 1U);
- }
- }
-
-
- /**
- *
- * @brief Process function for the Q15 Linear Interpolation Function.
- * @param[in] pYData pointer to Q15 Linear Interpolation table
- * @param[in] x input sample to process
- * @param[in] nValues number of table values
- * @return y processed output sample.
- *
- * \par
- * Input sample x is in 12.20 format which contains 12 bits for table index and 20 bits for fractional part.
- * This function can support maximum of table size 2^12.
- *
- */
- __STATIC_FORCEINLINE q15_t riscv_linear_interp_q15(
- q15_t * pYData,
- q31_t x,
- uint32_t nValues)
- {
- q63_t y; /* output */
- q15_t y0, y1; /* Nearest output values */
- q31_t fract; /* fractional part */
- int32_t index; /* Index to read nearest output values */
-
- /* Input is in 12.20 format */
- /* 12 bits for the table index */
- /* Index value calculation */
- index = ((x & (int32_t)0xFFF00000) >> 20);
-
- if (index >= (int32_t)(nValues - 1))
- {
- return (pYData[nValues - 1]);
- }
- else if (index < 0)
- {
- return (pYData[0]);
- }
- else
- {
- /* 20 bits for the fractional part */
- /* fract is in 12.20 format */
- fract = (x & 0x000FFFFF);
-
- /* Read two nearest output values from the index */
- y0 = pYData[index];
- y1 = pYData[index + 1];
-
- /* Calculation of y0 * (1-fract) and y is in 13.35 format */
- y = ((q63_t) y0 * (0xFFFFF - fract));
-
- /* Calculation of (y0 * (1-fract) + y1 * fract) and y is in 13.35 format */
- y += ((q63_t) y1 * (fract));
-
- /* convert y to 1.15 format */
- return (q15_t) (y >> 20);
- }
- }
-
-
- /**
- *
- * @brief Process function for the Q7 Linear Interpolation Function.
- * @param[in] pYData pointer to Q7 Linear Interpolation table
- * @param[in] x input sample to process
- * @param[in] nValues number of table values
- * @return y processed output sample.
- *
- * \par
- * Input sample x is in 12.20 format which contains 12 bits for table index and 20 bits for fractional part.
- * This function can support maximum of table size 2^12.
- */
- __STATIC_FORCEINLINE q7_t riscv_linear_interp_q7(
- q7_t * pYData,
- q31_t x,
- uint32_t nValues)
- {
- q31_t y; /* output */
- q7_t y0, y1; /* Nearest output values */
- q31_t fract; /* fractional part */
- uint32_t index; /* Index to read nearest output values */
-
- /* Input is in 12.20 format */
- /* 12 bits for the table index */
- /* Index value calculation */
- if (x < 0)
- {
- return (pYData[0]);
- }
- index = (x >> 20) & 0xfff;
-
- if (index >= (nValues - 1))
- {
- return (pYData[nValues - 1]);
- }
- else
- {
- /* 20 bits for the fractional part */
- /* fract is in 12.20 format */
- fract = (x & 0x000FFFFF);
-
- /* Read two nearest output values from the index and are in 1.7(q7) format */
- y0 = pYData[index];
- y1 = pYData[index + 1];
-
- /* Calculation of y0 * (1-fract ) and y is in 13.27(q27) format */
- y = ((y0 * (0xFFFFF - fract)));
-
- /* Calculation of y1 * fract + y0 * (1-fract) and y is in 13.27(q27) format */
- y += (y1 * fract);
-
- /* convert y to 1.7(q7) format */
- return (q7_t) (y >> 20);
- }
- }
-
- /**
- * @} end of LinearInterpolate group
- */
-
- /**
- * @brief Fast approximation to the trigonometric sine function for floating-point data.
- * @param[in] x input value in radians.
- * @return sin(x).
- */
- float32_t riscv_sin_f32(
- float32_t x);
-
-
- /**
- * @brief Fast approximation to the trigonometric sine function for Q31 data.
- * @param[in] x Scaled input value in radians.
- * @return sin(x).
- */
- q31_t riscv_sin_q31(
- q31_t x);
-
-
- /**
- * @brief Fast approximation to the trigonometric sine function for Q15 data.
- * @param[in] x Scaled input value in radians.
- * @return sin(x).
- */
- q15_t riscv_sin_q15(
- q15_t x);
-
-
- /**
- * @brief Fast approximation to the trigonometric cosine function for floating-point data.
- * @param[in] x input value in radians.
- * @return cos(x).
- */
- float32_t riscv_cos_f32(
- float32_t x);
-
-
- /**
- * @brief Fast approximation to the trigonometric cosine function for Q31 data.
- * @param[in] x Scaled input value in radians.
- * @return cos(x).
- */
- q31_t riscv_cos_q31(
- q31_t x);
-
-
- /**
- * @brief Fast approximation to the trigonometric cosine function for Q15 data.
- * @param[in] x Scaled input value in radians.
- * @return cos(x).
- */
- q15_t riscv_cos_q15(
- q15_t x);
-
-
- /**
- * @ingroup groupFastMath
- */
-
-
- /**
- * @defgroup SQRT Square Root
- *
- * Computes the square root of a number.
- * There are separate functions for Q15, Q31, and floating-point data types.
- * The square root function is computed using the Newton-Raphson algorithm.
- * This is an iterative algorithm of the form:
- *
- * x1 = x0 - f(x0)/f'(x0)
- *
- * where x1 is the current estimate,
- * x0 is the previous estimate, and
- * f'(x0) is the derivative of f() evaluated at x0.
- * For the square root function, the algorithm reduces to:
- *
- * x0 = in/2 [initial guess]
- * x1 = 1/2 * ( x0 + in / x0) [each iteration]
- *
- */
-
-
- /**
- * @addtogroup SQRT
- * @{
- */
-
-/**
- @brief Floating-point square root function.
- @param[in] in input value
- @param[out] pOut square root of input value
- @return execution status
- - \ref RISCV_MATH_SUCCESS : input value is positive
- - \ref RISCV_MATH_ARGUMENT_ERROR : input value is negative; *pOut is set to 0
- */
-__STATIC_FORCEINLINE riscv_status riscv_sqrt_f32(
- float32_t in,
- float32_t * pOut)
- {
-
- if (in >= 0.0f)
- {
-#if defined ( __riscv_flen )
- __ASM volatile("fsqrt.s %0, %1" : "=f"(*pOut) : "f"(in));
-#else
- *pOut = sqrtf(in);
-#endif /*__riscv_flen*/
-
- return (RISCV_MATH_SUCCESS);
- }
- else
- {
- *pOut = 0.0f;
- return (RISCV_MATH_ARGUMENT_ERROR);
- }
- }
-
-
-/**
- @brief Q31 square root function.
- @param[in] in input value. The range of the input value is [0 +1) or 0x00000000 to 0x7FFFFFFF
- @param[out] pOut points to square root of input value
- @return execution status
- - \ref RISCV_MATH_SUCCESS : input value is positive
- - \ref RISCV_MATH_ARGUMENT_ERROR : input value is negative; *pOut is set to 0
- */
-riscv_status riscv_sqrt_q31(
- q31_t in,
- q31_t * pOut);
-
-
-/**
- @brief Q15 square root function.
- @param[in] in input value. The range of the input value is [0 +1) or 0x0000 to 0x7FFF
- @param[out] pOut points to square root of input value
- @return execution status
- - \ref RISCV_MATH_SUCCESS : input value is positive
- - \ref RISCV_MATH_ARGUMENT_ERROR : input value is negative; *pOut is set to 0
- */
-riscv_status riscv_sqrt_q15(
- q15_t in,
- q15_t * pOut);
-
- /**
- * @brief Vector Floating-point square root function.
- * @param[in] pIn input vector.
- * @param[out] pOut vector of square roots of input elements.
- * @param[in] len length of input vector.
- * @return The function returns RISCV_MATH_SUCCESS if input value is positive value or RISCV_MATH_ARGUMENT_ERROR if
- * in is negative value and returns zero output for negative values.
- */
- void riscv_vsqrt_f32(
- float32_t * pIn,
- float32_t * pOut,
- uint16_t len);
-
- void riscv_vsqrt_q31(
- q31_t * pIn,
- q31_t * pOut,
- uint16_t len);
-
- void riscv_vsqrt_q15(
- q15_t * pIn,
- q15_t * pOut,
- uint16_t len);
-
- /**
- * @} end of SQRT group
- */
-
-
- /**
- * @brief floating-point Circular write function.
- */
- __STATIC_FORCEINLINE void riscv_circularWrite_f32(
- int32_t * circBuffer,
- int32_t L,
- uint16_t * writeOffset,
- int32_t bufferInc,
- const int32_t * src,
- int32_t srcInc,
- uint32_t blockSize)
- {
- uint32_t i = 0U;
- int32_t wOffset;
-
- /* Copy the value of Index pointer that points
- * to the current location where the input samples to be copied */
- wOffset = *writeOffset;
-
- /* Loop over the blockSize */
- i = blockSize;
-
- while (i > 0U)
- {
- /* copy the input sample to the circular buffer */
- circBuffer[wOffset] = *src;
-
- /* Update the input pointer */
- src += srcInc;
-
- /* Circularly update wOffset. Watch out for positive and negative value */
- wOffset += bufferInc;
- if (wOffset >= L)
- wOffset -= L;
-
- /* Decrement the loop counter */
- i--;
- }
-
- /* Update the index pointer */
- *writeOffset = (uint16_t)wOffset;
- }
-
-
-
- /**
- * @brief floating-point Circular Read function.
- */
- __STATIC_FORCEINLINE void riscv_circularRead_f32(
- int32_t * circBuffer,
- int32_t L,
- int32_t * readOffset,
- int32_t bufferInc,
- int32_t * dst,
- int32_t * dst_base,
- int32_t dst_length,
- int32_t dstInc,
- uint32_t blockSize)
- {
- uint32_t i = 0U;
- int32_t rOffset;
- int32_t* dst_end;
-
- /* Copy the value of Index pointer that points
- * to the current location from where the input samples to be read */
- rOffset = *readOffset;
- dst_end = dst_base + dst_length;
-
- /* Loop over the blockSize */
- i = blockSize;
-
- while (i > 0U)
- {
- /* copy the sample from the circular buffer to the destination buffer */
- *dst = circBuffer[rOffset];
-
- /* Update the input pointer */
- dst += dstInc;
-
- if (dst == dst_end)
- {
- dst = dst_base;
- }
-
- /* Circularly update rOffset. Watch out for positive and negative value */
- rOffset += bufferInc;
-
- if (rOffset >= L)
- {
- rOffset -= L;
- }
-
- /* Decrement the loop counter */
- i--;
- }
-
- /* Update the index pointer */
- *readOffset = rOffset;
- }
-
-
- /**
- * @brief Q15 Circular write function.
- */
- __STATIC_FORCEINLINE void riscv_circularWrite_q15(
- q15_t * circBuffer,
- int32_t L,
- uint16_t * writeOffset,
- int32_t bufferInc,
- const q15_t * src,
- int32_t srcInc,
- uint32_t blockSize)
- {
- uint32_t i = 0U;
- int32_t wOffset;
-
- /* Copy the value of Index pointer that points
- * to the current location where the input samples to be copied */
- wOffset = *writeOffset;
-
- /* Loop over the blockSize */
- i = blockSize;
-
- while (i > 0U)
- {
- /* copy the input sample to the circular buffer */
- circBuffer[wOffset] = *src;
-
- /* Update the input pointer */
- src += srcInc;
-
- /* Circularly update wOffset. Watch out for positive and negative value */
- wOffset += bufferInc;
- if (wOffset >= L)
- wOffset -= L;
-
- /* Decrement the loop counter */
- i--;
- }
-
- /* Update the index pointer */
- *writeOffset = (uint16_t)wOffset;
- }
-
-
- /**
- * @brief Q15 Circular Read function.
- */
- __STATIC_FORCEINLINE void riscv_circularRead_q15(
- q15_t * circBuffer,
- int32_t L,
- int32_t * readOffset,
- int32_t bufferInc,
- q15_t * dst,
- q15_t * dst_base,
- int32_t dst_length,
- int32_t dstInc,
- uint32_t blockSize)
- {
- uint32_t i = 0;
- int32_t rOffset;
- q15_t* dst_end;
-
- /* Copy the value of Index pointer that points
- * to the current location from where the input samples to be read */
- rOffset = *readOffset;
-
- dst_end = dst_base + dst_length;
-
- /* Loop over the blockSize */
- i = blockSize;
-
- while (i > 0U)
- {
- /* copy the sample from the circular buffer to the destination buffer */
- *dst = circBuffer[rOffset];
-
- /* Update the input pointer */
- dst += dstInc;
-
- if (dst == dst_end)
- {
- dst = dst_base;
- }
-
- /* Circularly update wOffset. Watch out for positive and negative value */
- rOffset += bufferInc;
-
- if (rOffset >= L)
- {
- rOffset -= L;
- }
-
- /* Decrement the loop counter */
- i--;
- }
-
- /* Update the index pointer */
- *readOffset = rOffset;
- }
-
-
- /**
- * @brief Q7 Circular write function.
- */
- __STATIC_FORCEINLINE void riscv_circularWrite_q7(
- q7_t * circBuffer,
- int32_t L,
- uint16_t * writeOffset,
- int32_t bufferInc,
- const q7_t * src,
- int32_t srcInc,
- uint32_t blockSize)
- {
- uint32_t i = 0U;
- int32_t wOffset;
-
- /* Copy the value of Index pointer that points
- * to the current location where the input samples to be copied */
- wOffset = *writeOffset;
-
- /* Loop over the blockSize */
- i = blockSize;
-
- while (i > 0U)
- {
- /* copy the input sample to the circular buffer */
- circBuffer[wOffset] = *src;
-
- /* Update the input pointer */
- src += srcInc;
-
- /* Circularly update wOffset. Watch out for positive and negative value */
- wOffset += bufferInc;
- if (wOffset >= L)
- wOffset -= L;
-
- /* Decrement the loop counter */
- i--;
- }
-
- /* Update the index pointer */
- *writeOffset = (uint16_t)wOffset;
- }
-
-
- /**
- * @brief Q7 Circular Read function.
- */
- __STATIC_FORCEINLINE void riscv_circularRead_q7(
- q7_t * circBuffer,
- int32_t L,
- int32_t * readOffset,
- int32_t bufferInc,
- q7_t * dst,
- q7_t * dst_base,
- int32_t dst_length,
- int32_t dstInc,
- uint32_t blockSize)
- {
- uint32_t i = 0;
- int32_t rOffset;
- q7_t* dst_end;
-
- /* Copy the value of Index pointer that points
- * to the current location from where the input samples to be read */
- rOffset = *readOffset;
-
- dst_end = dst_base + dst_length;
-
- /* Loop over the blockSize */
- i = blockSize;
-
- while (i > 0U)
- {
- /* copy the sample from the circular buffer to the destination buffer */
- *dst = circBuffer[rOffset];
-
- /* Update the input pointer */
- dst += dstInc;
-
- if (dst == dst_end)
- {
- dst = dst_base;
- }
-
- /* Circularly update rOffset. Watch out for positive and negative value */
- rOffset += bufferInc;
-
- if (rOffset >= L)
- {
- rOffset -= L;
- }
-
- /* Decrement the loop counter */
- i--;
- }
-
- /* Update the index pointer */
- *readOffset = rOffset;
- }
-
-
- /**
- * @brief Sum of the squares of the elements of a Q31 vector.
- * @param[in] pSrc is input pointer
- * @param[in] blockSize is the number of samples to process
- * @param[out] pResult is output value.
- */
- void riscv_power_q31(
- const q31_t * pSrc,
- uint32_t blockSize,
- q63_t * pResult);
-
-
- /**
- * @brief Sum of the squares of the elements of a floating-point vector.
- * @param[in] pSrc is input pointer
- * @param[in] blockSize is the number of samples to process
- * @param[out] pResult is output value.
- */
- void riscv_power_f32(
- const float32_t * pSrc,
- uint32_t blockSize,
- float32_t * pResult);
-
-
- /**
- * @brief Sum of the squares of the elements of a Q15 vector.
- * @param[in] pSrc is input pointer
- * @param[in] blockSize is the number of samples to process
- * @param[out] pResult is output value.
- */
- void riscv_power_q15(
- const q15_t * pSrc,
- uint32_t blockSize,
- q63_t * pResult);
-
-
- /**
- * @brief Sum of the squares of the elements of a Q7 vector.
- * @param[in] pSrc is input pointer
- * @param[in] blockSize is the number of samples to process
- * @param[out] pResult is output value.
- */
- void riscv_power_q7(
- const q7_t * pSrc,
- uint32_t blockSize,
- q31_t * pResult);
-
-
- /**
- * @brief Mean value of a Q7 vector.
- * @param[in] pSrc is input pointer
- * @param[in] blockSize is the number of samples to process
- * @param[out] pResult is output value.
- */
- void riscv_mean_q7(
- const q7_t * pSrc,
- uint32_t blockSize,
- q7_t * pResult);
-
-
- /**
- * @brief Mean value of a Q15 vector.
- * @param[in] pSrc is input pointer
- * @param[in] blockSize is the number of samples to process
- * @param[out] pResult is output value.
- */
- void riscv_mean_q15(
- const q15_t * pSrc,
- uint32_t blockSize,
- q15_t * pResult);
-
-
- /**
- * @brief Mean value of a Q31 vector.
- * @param[in] pSrc is input pointer
- * @param[in] blockSize is the number of samples to process
- * @param[out] pResult is output value.
- */
- void riscv_mean_q31(
- const q31_t * pSrc,
- uint32_t blockSize,
- q31_t * pResult);
-
-
- /**
- * @brief Mean value of a floating-point vector.
- * @param[in] pSrc is input pointer
- * @param[in] blockSize is the number of samples to process
- * @param[out] pResult is output value.
- */
- void riscv_mean_f32(
- const float32_t * pSrc,
- uint32_t blockSize,
- float32_t * pResult);
-
-
- /**
- * @brief Variance of the elements of a floating-point vector.
- * @param[in] pSrc is input pointer
- * @param[in] blockSize is the number of samples to process
- * @param[out] pResult is output value.
- */
- void riscv_var_f32(
- const float32_t * pSrc,
- uint32_t blockSize,
- float32_t * pResult);
-
-
- /**
- * @brief Variance of the elements of a Q31 vector.
- * @param[in] pSrc is input pointer
- * @param[in] blockSize is the number of samples to process
- * @param[out] pResult is output value.
- */
- void riscv_var_q31(
- const q31_t * pSrc,
- uint32_t blockSize,
- q31_t * pResult);
-
-
- /**
- * @brief Variance of the elements of a Q15 vector.
- * @param[in] pSrc is input pointer
- * @param[in] blockSize is the number of samples to process
- * @param[out] pResult is output value.
- */
- void riscv_var_q15(
- const q15_t * pSrc,
- uint32_t blockSize,
- q15_t * pResult);
-
-
- /**
- * @brief Root Mean Square of the elements of a floating-point vector.
- * @param[in] pSrc is input pointer
- * @param[in] blockSize is the number of samples to process
- * @param[out] pResult is output value.
- */
- void riscv_rms_f32(
- const float32_t * pSrc,
- uint32_t blockSize,
- float32_t * pResult);
-
-
- /**
- * @brief Root Mean Square of the elements of a Q31 vector.
- * @param[in] pSrc is input pointer
- * @param[in] blockSize is the number of samples to process
- * @param[out] pResult is output value.
- */
- void riscv_rms_q31(
- const q31_t * pSrc,
- uint32_t blockSize,
- q31_t * pResult);
-
-
- /**
- * @brief Root Mean Square of the elements of a Q15 vector.
- * @param[in] pSrc is input pointer
- * @param[in] blockSize is the number of samples to process
- * @param[out] pResult is output value.
- */
- void riscv_rms_q15(
- const q15_t * pSrc,
- uint32_t blockSize,
- q15_t * pResult);
-
-
- /**
- * @brief Standard deviation of the elements of a floating-point vector.
- * @param[in] pSrc is input pointer
- * @param[in] blockSize is the number of samples to process
- * @param[out] pResult is output value.
- */
- void riscv_std_f32(
- const float32_t * pSrc,
- uint32_t blockSize,
- float32_t * pResult);
-
-
- /**
- * @brief Standard deviation of the elements of a Q31 vector.
- * @param[in] pSrc is input pointer
- * @param[in] blockSize is the number of samples to process
- * @param[out] pResult is output value.
- */
- void riscv_std_q31(
- const q31_t * pSrc,
- uint32_t blockSize,
- q31_t * pResult);
-
-
- /**
- * @brief Standard deviation of the elements of a Q15 vector.
- * @param[in] pSrc is input pointer
- * @param[in] blockSize is the number of samples to process
- * @param[out] pResult is output value.
- */
- void riscv_std_q15(
- const q15_t * pSrc,
- uint32_t blockSize,
- q15_t * pResult);
-
-
- /**
- * @brief Floating-point complex magnitude
- * @param[in] pSrc points to the complex input vector
- * @param[out] pDst points to the real output vector
- * @param[in] numSamples number of complex samples in the input vector
- */
- void riscv_cmplx_mag_f32(
- const float32_t * pSrc,
- float32_t * pDst,
- uint32_t numSamples);
-
-
- /**
- * @brief Q31 complex magnitude
- * @param[in] pSrc points to the complex input vector
- * @param[out] pDst points to the real output vector
- * @param[in] numSamples number of complex samples in the input vector
- */
- void riscv_cmplx_mag_q31(
- const q31_t * pSrc,
- q31_t * pDst,
- uint32_t numSamples);
-
-
- /**
- * @brief Q15 complex magnitude
- * @param[in] pSrc points to the complex input vector
- * @param[out] pDst points to the real output vector
- * @param[in] numSamples number of complex samples in the input vector
- */
- void riscv_cmplx_mag_q15(
- const q15_t * pSrc,
- q15_t * pDst,
- uint32_t numSamples);
-
-
- /**
- * @brief Q15 complex dot product
- * @param[in] pSrcA points to the first input vector
- * @param[in] pSrcB points to the second input vector
- * @param[in] numSamples number of complex samples in each vector
- * @param[out] realResult real part of the result returned here
- * @param[out] imagResult imaginary part of the result returned here
- */
- void riscv_cmplx_dot_prod_q15(
- const q15_t * pSrcA,
- const q15_t * pSrcB,
- uint32_t numSamples,
- q31_t * realResult,
- q31_t * imagResult);
-
-
- /**
- * @brief Q31 complex dot product
- * @param[in] pSrcA points to the first input vector
- * @param[in] pSrcB points to the second input vector
- * @param[in] numSamples number of complex samples in each vector
- * @param[out] realResult real part of the result returned here
- * @param[out] imagResult imaginary part of the result returned here
- */
- void riscv_cmplx_dot_prod_q31(
- const q31_t * pSrcA,
- const q31_t * pSrcB,
- uint32_t numSamples,
- q63_t * realResult,
- q63_t * imagResult);
-
-
- /**
- * @brief Floating-point complex dot product
- * @param[in] pSrcA points to the first input vector
- * @param[in] pSrcB points to the second input vector
- * @param[in] numSamples number of complex samples in each vector
- * @param[out] realResult real part of the result returned here
- * @param[out] imagResult imaginary part of the result returned here
- */
- void riscv_cmplx_dot_prod_f32(
- const float32_t * pSrcA,
- const float32_t * pSrcB,
- uint32_t numSamples,
- float32_t * realResult,
- float32_t * imagResult);
-
-
- /**
- * @brief Q15 complex-by-real multiplication
- * @param[in] pSrcCmplx points to the complex input vector
- * @param[in] pSrcReal points to the real input vector
- * @param[out] pCmplxDst points to the complex output vector
- * @param[in] numSamples number of samples in each vector
- */
- void riscv_cmplx_mult_real_q15(
- const q15_t * pSrcCmplx,
- const q15_t * pSrcReal,
- q15_t * pCmplxDst,
- uint32_t numSamples);
-
-
- /**
- * @brief Q31 complex-by-real multiplication
- * @param[in] pSrcCmplx points to the complex input vector
- * @param[in] pSrcReal points to the real input vector
- * @param[out] pCmplxDst points to the complex output vector
- * @param[in] numSamples number of samples in each vector
- */
- void riscv_cmplx_mult_real_q31(
- const q31_t * pSrcCmplx,
- const q31_t * pSrcReal,
- q31_t * pCmplxDst,
- uint32_t numSamples);
-
-
- /**
- * @brief Floating-point complex-by-real multiplication
- * @param[in] pSrcCmplx points to the complex input vector
- * @param[in] pSrcReal points to the real input vector
- * @param[out] pCmplxDst points to the complex output vector
- * @param[in] numSamples number of samples in each vector
- */
- void riscv_cmplx_mult_real_f32(
- const float32_t * pSrcCmplx,
- const float32_t * pSrcReal,
- float32_t * pCmplxDst,
- uint32_t numSamples);
-
-
- /**
- * @brief Minimum value of a Q7 vector.
- * @param[in] pSrc is input pointer
- * @param[in] blockSize is the number of samples to process
- * @param[out] result is output pointer
- * @param[in] index is the array index of the minimum value in the input buffer.
- */
- void riscv_min_q7(
- const q7_t * pSrc,
- uint32_t blockSize,
- q7_t * result,
- uint32_t * index);
-
-
- /**
- * @brief Minimum value of a Q15 vector.
- * @param[in] pSrc is input pointer
- * @param[in] blockSize is the number of samples to process
- * @param[out] pResult is output pointer
- * @param[in] pIndex is the array index of the minimum value in the input buffer.
- */
- void riscv_min_q15(
- const q15_t * pSrc,
- uint32_t blockSize,
- q15_t * pResult,
- uint32_t * pIndex);
-
-
- /**
- * @brief Minimum value of a Q31 vector.
- * @param[in] pSrc is input pointer
- * @param[in] blockSize is the number of samples to process
- * @param[out] pResult is output pointer
- * @param[out] pIndex is the array index of the minimum value in the input buffer.
- */
- void riscv_min_q31(
- const q31_t * pSrc,
- uint32_t blockSize,
- q31_t * pResult,
- uint32_t * pIndex);
-
-
- /**
- * @brief Minimum value of a floating-point vector.
- * @param[in] pSrc is input pointer
- * @param[in] blockSize is the number of samples to process
- * @param[out] pResult is output pointer
- * @param[out] pIndex is the array index of the minimum value in the input buffer.
- */
- void riscv_min_f32(
- const float32_t * pSrc,
- uint32_t blockSize,
- float32_t * pResult,
- uint32_t * pIndex);
-
-
-/**
- * @brief Maximum value of a Q7 vector.
- * @param[in] pSrc points to the input buffer
- * @param[in] blockSize length of the input vector
- * @param[out] pResult maximum value returned here
- * @param[out] pIndex index of maximum value returned here
- */
- void riscv_max_q7(
- const q7_t * pSrc,
- uint32_t blockSize,
- q7_t * pResult,
- uint32_t * pIndex);
-
-
-/**
- * @brief Maximum value of a Q15 vector.
- * @param[in] pSrc points to the input buffer
- * @param[in] blockSize length of the input vector
- * @param[out] pResult maximum value returned here
- * @param[out] pIndex index of maximum value returned here
- */
- void riscv_max_q15(
- const q15_t * pSrc,
- uint32_t blockSize,
- q15_t * pResult,
- uint32_t * pIndex);
-
-
-/**
- * @brief Maximum value of a Q31 vector.
- * @param[in] pSrc points to the input buffer
- * @param[in] blockSize length of the input vector
- * @param[out] pResult maximum value returned here
- * @param[out] pIndex index of maximum value returned here
- */
- void riscv_max_q31(
- const q31_t * pSrc,
- uint32_t blockSize,
- q31_t * pResult,
- uint32_t * pIndex);
-
-
-/**
- * @brief Maximum value of a floating-point vector.
- * @param[in] pSrc points to the input buffer
- * @param[in] blockSize length of the input vector
- * @param[out] pResult maximum value returned here
- * @param[out] pIndex index of maximum value returned here
- */
- void riscv_max_f32(
- const float32_t * pSrc,
- uint32_t blockSize,
- float32_t * pResult,
- uint32_t * pIndex);
-
-
- /**
- * @brief Q15 complex-by-complex multiplication
- * @param[in] pSrcA points to the first input vector
- * @param[in] pSrcB points to the second input vector
- * @param[out] pDst points to the output vector
- * @param[in] numSamples number of complex samples in each vector
- */
- void riscv_cmplx_mult_cmplx_q15(
- const q15_t * pSrcA,
- const q15_t * pSrcB,
- q15_t * pDst,
- uint32_t numSamples);
-
-
- /**
- * @brief Q31 complex-by-complex multiplication
- * @param[in] pSrcA points to the first input vector
- * @param[in] pSrcB points to the second input vector
- * @param[out] pDst points to the output vector
- * @param[in] numSamples number of complex samples in each vector
- */
- void riscv_cmplx_mult_cmplx_q31(
- const q31_t * pSrcA,
- const q31_t * pSrcB,
- q31_t * pDst,
- uint32_t numSamples);
-
-
- /**
- * @brief Floating-point complex-by-complex multiplication
- * @param[in] pSrcA points to the first input vector
- * @param[in] pSrcB points to the second input vector
- * @param[out] pDst points to the output vector
- * @param[in] numSamples number of complex samples in each vector
- */
- void riscv_cmplx_mult_cmplx_f32(
- const float32_t * pSrcA,
- const float32_t * pSrcB,
- float32_t * pDst,
- uint32_t numSamples);
-
-
- /**
- * @brief Converts the elements of the floating-point vector to Q31 vector.
- * @param[in] pSrc points to the floating-point input vector
- * @param[out] pDst points to the Q31 output vector
- * @param[in] blockSize length of the input vector
- */
- void riscv_float_to_q31(
- const float32_t * pSrc,
- q31_t * pDst,
- uint32_t blockSize);
-
-
- /**
- * @brief Converts the elements of the floating-point vector to Q15 vector.
- * @param[in] pSrc points to the floating-point input vector
- * @param[out] pDst points to the Q15 output vector
- * @param[in] blockSize length of the input vector
- */
- void riscv_float_to_q15(
- const float32_t * pSrc,
- q15_t * pDst,
- uint32_t blockSize);
-
-
- /**
- * @brief Converts the elements of the floating-point vector to Q7 vector.
- * @param[in] pSrc points to the floating-point input vector
- * @param[out] pDst points to the Q7 output vector
- * @param[in] blockSize length of the input vector
- */
- void riscv_float_to_q7(
- const float32_t * pSrc,
- q7_t * pDst,
- uint32_t blockSize);
-
-
- /**
- * @brief Converts the elements of the Q31 vector to floating-point vector.
- * @param[in] pSrc is input pointer
- * @param[out] pDst is output pointer
- * @param[in] blockSize is the number of samples to process
- */
- void riscv_q31_to_float(
- const q31_t * pSrc,
- float32_t * pDst,
- uint32_t blockSize);
-
-
- /**
- * @brief Converts the elements of the Q31 vector to Q15 vector.
- * @param[in] pSrc is input pointer
- * @param[out] pDst is output pointer
- * @param[in] blockSize is the number of samples to process
- */
- void riscv_q31_to_q15(
- const q31_t * pSrc,
- q15_t * pDst,
- uint32_t blockSize);
-
-
- /**
- * @brief Converts the elements of the Q31 vector to Q7 vector.
- * @param[in] pSrc is input pointer
- * @param[out] pDst is output pointer
- * @param[in] blockSize is the number of samples to process
- */
- void riscv_q31_to_q7(
- const q31_t * pSrc,
- q7_t * pDst,
- uint32_t blockSize);
-
-
- /**
- * @brief Converts the elements of the Q15 vector to floating-point vector.
- * @param[in] pSrc is input pointer
- * @param[out] pDst is output pointer
- * @param[in] blockSize is the number of samples to process
- */
- void riscv_q15_to_float(
- const q15_t * pSrc,
- float32_t * pDst,
- uint32_t blockSize);
-
-
- /**
- * @brief Converts the elements of the Q15 vector to Q31 vector.
- * @param[in] pSrc is input pointer
- * @param[out] pDst is output pointer
- * @param[in] blockSize is the number of samples to process
- */
- void riscv_q15_to_q31(
- const q15_t * pSrc,
- q31_t * pDst,
- uint32_t blockSize);
-
-
- /**
- * @brief Converts the elements of the Q15 vector to Q7 vector.
- * @param[in] pSrc is input pointer
- * @param[out] pDst is output pointer
- * @param[in] blockSize is the number of samples to process
- */
- void riscv_q15_to_q7(
- const q15_t * pSrc,
- q7_t * pDst,
- uint32_t blockSize);
-
-
- /**
- * @brief Converts the elements of the Q7 vector to floating-point vector.
- * @param[in] pSrc is input pointer
- * @param[out] pDst is output pointer
- * @param[in] blockSize is the number of samples to process
- */
- void riscv_q7_to_float(
- const q7_t * pSrc,
- float32_t * pDst,
- uint32_t blockSize);
-
-
- /**
- * @brief Converts the elements of the Q7 vector to Q31 vector.
- * @param[in] pSrc input pointer
- * @param[out] pDst output pointer
- * @param[in] blockSize number of samples to process
- */
- void riscv_q7_to_q31(
- const q7_t * pSrc,
- q31_t * pDst,
- uint32_t blockSize);
-
-
- /**
- * @brief Converts the elements of the Q7 vector to Q15 vector.
- * @param[in] pSrc input pointer
- * @param[out] pDst output pointer
- * @param[in] blockSize number of samples to process
- */
- void riscv_q7_to_q15(
- const q7_t * pSrc,
- q15_t * pDst,
- uint32_t blockSize);
-
-
- /**
- * @ingroup groupInterpolation
- */
-
- /**
- * @defgroup BilinearInterpolate Bilinear Interpolation
- *
- * Bilinear interpolation is an extension of linear interpolation applied to a two dimensional grid.
- * The underlying function f(x, y) is sampled on a regular grid and the interpolation process
- * determines values between the grid points.
- * Bilinear interpolation is equivalent to two step linear interpolation, first in the x-dimension and then in the y-dimension.
- * Bilinear interpolation is often used in image processing to rescale images.
- * The NMSIS DSP library provides bilinear interpolation functions for Q7, Q15, Q31, and floating-point data types.
- *
- * Algorithm
- * \par
- * The instance structure used by the bilinear interpolation functions describes a two dimensional data table.
- * For floating-point, the instance structure is defined as:
- *
- * typedef struct
- * {
- * uint16_t numRows;
- * uint16_t numCols;
- * float32_t *pData;
- * } riscv_bilinear_interp_instance_f32;
- *
- *
- * \par
- * where numRows specifies the number of rows in the table;
- * numCols specifies the number of columns in the table;
- * and pData points to an array of size numRows*numCols values.
- * The data table pTable is organized in row order and the supplied data values fall on integer indexes.
- * That is, table element (x,y) is located at pTable[x + y*numCols] where x and y are integers.
- *
- * \par
- * Let (x, y) specify the desired interpolation point. Then define:
- *
- * XF = floor(x)
- * YF = floor(y)
- *
- * \par
- * The interpolated output point is computed as:
- *
- * f(x, y) = f(XF, YF) * (1-(x-XF)) * (1-(y-YF))
- * + f(XF+1, YF) * (x-XF)*(1-(y-YF))
- * + f(XF, YF+1) * (1-(x-XF))*(y-YF)
- * + f(XF+1, YF+1) * (x-XF)*(y-YF)
- *
- * Note that the coordinates (x, y) contain integer and fractional components.
- * The integer components specify which portion of the table to use while the
- * fractional components control the interpolation processor.
- *
- * \par
- * if (x,y) are outside of the table boundary, Bilinear interpolation returns zero output.
- */
-
-
- /**
- * @addtogroup BilinearInterpolate
- * @{
- */
-
- /**
- * @brief Floating-point bilinear interpolation.
- * @param[in,out] S points to an instance of the interpolation structure.
- * @param[in] X interpolation coordinate.
- * @param[in] Y interpolation coordinate.
- * @return out interpolated value.
- */
- __STATIC_FORCEINLINE float32_t riscv_bilinear_interp_f32(
- const riscv_bilinear_interp_instance_f32 * S,
- float32_t X,
- float32_t Y)
- {
- float32_t out;
- float32_t f00, f01, f10, f11;
- float32_t *pData = S->pData;
- int32_t xIndex, yIndex, index;
- float32_t xdiff, ydiff;
- float32_t b1, b2, b3, b4;
-
- xIndex = (int32_t) X;
- yIndex = (int32_t) Y;
-
- /* Care taken for table outside boundary */
- /* Returns zero output when values are outside table boundary */
- if (xIndex < 0 || xIndex > (S->numRows - 1) || yIndex < 0 || yIndex > (S->numCols - 1))
- {
- return (0);
- }
-
- /* Calculation of index for two nearest points in X-direction */
- index = (xIndex - 1) + (yIndex - 1) * S->numCols;
-
-
- /* Read two nearest points in X-direction */
- f00 = pData[index];
- f01 = pData[index + 1];
-
- /* Calculation of index for two nearest points in Y-direction */
- index = (xIndex - 1) + (yIndex) * S->numCols;
-
-
- /* Read two nearest points in Y-direction */
- f10 = pData[index];
- f11 = pData[index + 1];
-
- /* Calculation of intermediate values */
- b1 = f00;
- b2 = f01 - f00;
- b3 = f10 - f00;
- b4 = f00 - f01 - f10 + f11;
-
- /* Calculation of fractional part in X */
- xdiff = X - xIndex;
-
- /* Calculation of fractional part in Y */
- ydiff = Y - yIndex;
-
- /* Calculation of bi-linear interpolated output */
- out = b1 + b2 * xdiff + b3 * ydiff + b4 * xdiff * ydiff;
-
- /* return to application */
- return (out);
- }
-
-
- /**
- * @brief Q31 bilinear interpolation.
- * @param[in,out] S points to an instance of the interpolation structure.
- * @param[in] X interpolation coordinate in 12.20 format.
- * @param[in] Y interpolation coordinate in 12.20 format.
- * @return out interpolated value.
- */
- __STATIC_FORCEINLINE q31_t riscv_bilinear_interp_q31(
- riscv_bilinear_interp_instance_q31 * S,
- q31_t X,
- q31_t Y)
- {
- q31_t out; /* Temporary output */
- q31_t acc = 0; /* output */
- q31_t xfract, yfract; /* X, Y fractional parts */
- q31_t x1, x2, y1, y2; /* Nearest output values */
- int32_t rI, cI; /* Row and column indices */
- q31_t *pYData = S->pData; /* pointer to output table values */
- uint32_t nCols = S->numCols; /* num of rows */
-
- /* Input is in 12.20 format */
- /* 12 bits for the table index */
- /* Index value calculation */
- rI = ((X & (q31_t)0xFFF00000) >> 20);
-
- /* Input is in 12.20 format */
- /* 12 bits for the table index */
- /* Index value calculation */
- cI = ((Y & (q31_t)0xFFF00000) >> 20);
-
- /* Care taken for table outside boundary */
- /* Returns zero output when values are outside table boundary */
- if (rI < 0 || rI > (S->numRows - 1) || cI < 0 || cI > (S->numCols - 1))
- {
- return (0);
- }
-
- /* 20 bits for the fractional part */
- /* shift left xfract by 11 to keep 1.31 format */
- xfract = (X & 0x000FFFFF) << 11U;
-
- /* Read two nearest output values from the index */
- x1 = pYData[(rI) + (int32_t)nCols * (cI) ];
- x2 = pYData[(rI) + (int32_t)nCols * (cI) + 1];
-
- /* 20 bits for the fractional part */
- /* shift left yfract by 11 to keep 1.31 format */
- yfract = (Y & 0x000FFFFF) << 11U;
-
- /* Read two nearest output values from the index */
- y1 = pYData[(rI) + (int32_t)nCols * (cI + 1) ];
- y2 = pYData[(rI) + (int32_t)nCols * (cI + 1) + 1];
-
- /* Calculation of x1 * (1-xfract ) * (1-yfract) and acc is in 3.29(q29) format */
- out = ((q31_t) (((q63_t) x1 * (0x7FFFFFFF - xfract)) >> 32));
- acc = ((q31_t) (((q63_t) out * (0x7FFFFFFF - yfract)) >> 32));
-
- /* x2 * (xfract) * (1-yfract) in 3.29(q29) and adding to acc */
- out = ((q31_t) ((q63_t) x2 * (0x7FFFFFFF - yfract) >> 32));
- acc += ((q31_t) ((q63_t) out * (xfract) >> 32));
-
- /* y1 * (1 - xfract) * (yfract) in 3.29(q29) and adding to acc */
- out = ((q31_t) ((q63_t) y1 * (0x7FFFFFFF - xfract) >> 32));
- acc += ((q31_t) ((q63_t) out * (yfract) >> 32));
-
- /* y2 * (xfract) * (yfract) in 3.29(q29) and adding to acc */
- out = ((q31_t) ((q63_t) y2 * (xfract) >> 32));
- acc += ((q31_t) ((q63_t) out * (yfract) >> 32));
-
- /* Convert acc to 1.31(q31) format */
- return ((q31_t)(acc << 2));
- }
-
-
- /**
- * @brief Q15 bilinear interpolation.
- * @param[in,out] S points to an instance of the interpolation structure.
- * @param[in] X interpolation coordinate in 12.20 format.
- * @param[in] Y interpolation coordinate in 12.20 format.
- * @return out interpolated value.
- */
- __STATIC_FORCEINLINE q15_t riscv_bilinear_interp_q15(
- riscv_bilinear_interp_instance_q15 * S,
- q31_t X,
- q31_t Y)
- {
- q63_t acc = 0; /* output */
- q31_t out; /* Temporary output */
- q15_t x1, x2, y1, y2; /* Nearest output values */
- q31_t xfract, yfract; /* X, Y fractional parts */
- int32_t rI, cI; /* Row and column indices */
- q15_t *pYData = S->pData; /* pointer to output table values */
- uint32_t nCols = S->numCols; /* num of rows */
-
- /* Input is in 12.20 format */
- /* 12 bits for the table index */
- /* Index value calculation */
- rI = ((X & (q31_t)0xFFF00000) >> 20);
-
- /* Input is in 12.20 format */
- /* 12 bits for the table index */
- /* Index value calculation */
- cI = ((Y & (q31_t)0xFFF00000) >> 20);
-
- /* Care taken for table outside boundary */
- /* Returns zero output when values are outside table boundary */
- if (rI < 0 || rI > (S->numRows - 1) || cI < 0 || cI > (S->numCols - 1))
- {
- return (0);
- }
-
- /* 20 bits for the fractional part */
- /* xfract should be in 12.20 format */
- xfract = (X & 0x000FFFFF);
-
- /* Read two nearest output values from the index */
- x1 = pYData[((uint32_t)rI) + nCols * ((uint32_t)cI) ];
- x2 = pYData[((uint32_t)rI) + nCols * ((uint32_t)cI) + 1];
-
- /* 20 bits for the fractional part */
- /* yfract should be in 12.20 format */
- yfract = (Y & 0x000FFFFF);
-
- /* Read two nearest output values from the index */
- y1 = pYData[((uint32_t)rI) + nCols * ((uint32_t)cI + 1) ];
- y2 = pYData[((uint32_t)rI) + nCols * ((uint32_t)cI + 1) + 1];
-
- /* Calculation of x1 * (1-xfract ) * (1-yfract) and acc is in 13.51 format */
-
- /* x1 is in 1.15(q15), xfract in 12.20 format and out is in 13.35 format */
- /* convert 13.35 to 13.31 by right shifting and out is in 1.31 */
- out = (q31_t) (((q63_t) x1 * (0xFFFFF - xfract)) >> 4U);
- acc = ((q63_t) out * (0xFFFFF - yfract));
-
- /* x2 * (xfract) * (1-yfract) in 1.51 and adding to acc */
- out = (q31_t) (((q63_t) x2 * (0xFFFFF - yfract)) >> 4U);
- acc += ((q63_t) out * (xfract));
-
- /* y1 * (1 - xfract) * (yfract) in 1.51 and adding to acc */
- out = (q31_t) (((q63_t) y1 * (0xFFFFF - xfract)) >> 4U);
- acc += ((q63_t) out * (yfract));
-
- /* y2 * (xfract) * (yfract) in 1.51 and adding to acc */
- out = (q31_t) (((q63_t) y2 * (xfract)) >> 4U);
- acc += ((q63_t) out * (yfract));
-
- /* acc is in 13.51 format and down shift acc by 36 times */
- /* Convert out to 1.15 format */
- return ((q15_t)(acc >> 36));
- }
-
-
- /**
- * @brief Q7 bilinear interpolation.
- * @param[in,out] S points to an instance of the interpolation structure.
- * @param[in] X interpolation coordinate in 12.20 format.
- * @param[in] Y interpolation coordinate in 12.20 format.
- * @return out interpolated value.
- */
- __STATIC_FORCEINLINE q7_t riscv_bilinear_interp_q7(
- riscv_bilinear_interp_instance_q7 * S,
- q31_t X,
- q31_t Y)
- {
- q63_t acc = 0; /* output */
- q31_t out; /* Temporary output */
- q31_t xfract, yfract; /* X, Y fractional parts */
- q7_t x1, x2, y1, y2; /* Nearest output values */
- int32_t rI, cI; /* Row and column indices */
- q7_t *pYData = S->pData; /* pointer to output table values */
- uint32_t nCols = S->numCols; /* num of rows */
-
- /* Input is in 12.20 format */
- /* 12 bits for the table index */
- /* Index value calculation */
- rI = ((X & (q31_t)0xFFF00000) >> 20);
-
- /* Input is in 12.20 format */
- /* 12 bits for the table index */
- /* Index value calculation */
- cI = ((Y & (q31_t)0xFFF00000) >> 20);
-
- /* Care taken for table outside boundary */
- /* Returns zero output when values are outside table boundary */
- if (rI < 0 || rI > (S->numRows - 1) || cI < 0 || cI > (S->numCols - 1))
- {
- return (0);
- }
-
- /* 20 bits for the fractional part */
- /* xfract should be in 12.20 format */
- xfract = (X & (q31_t)0x000FFFFF);
-
- /* Read two nearest output values from the index */
- x1 = pYData[((uint32_t)rI) + nCols * ((uint32_t)cI) ];
- x2 = pYData[((uint32_t)rI) + nCols * ((uint32_t)cI) + 1];
-
- /* 20 bits for the fractional part */
- /* yfract should be in 12.20 format */
- yfract = (Y & (q31_t)0x000FFFFF);
-
- /* Read two nearest output values from the index */
- y1 = pYData[((uint32_t)rI) + nCols * ((uint32_t)cI + 1) ];
- y2 = pYData[((uint32_t)rI) + nCols * ((uint32_t)cI + 1) + 1];
-
- /* Calculation of x1 * (1-xfract ) * (1-yfract) and acc is in 16.47 format */
- out = ((x1 * (0xFFFFF - xfract)));
- acc = (((q63_t) out * (0xFFFFF - yfract)));
-
- /* x2 * (xfract) * (1-yfract) in 2.22 and adding to acc */
- out = ((x2 * (0xFFFFF - yfract)));
- acc += (((q63_t) out * (xfract)));
-
- /* y1 * (1 - xfract) * (yfract) in 2.22 and adding to acc */
- out = ((y1 * (0xFFFFF - xfract)));
- acc += (((q63_t) out * (yfract)));
-
- /* y2 * (xfract) * (yfract) in 2.22 and adding to acc */
- out = ((y2 * (yfract)));
- acc += (((q63_t) out * (xfract)));
-
- /* acc in 16.47 format and down shift by 40 to convert to 1.7 format */
- return ((q7_t)(acc >> 40));
- }
-
- /**
- * @} end of BilinearInterpolate group
- */
-
-
-/* SMMLAR */
-#define multAcc_32x32_keep32_R(a, x, y) \
- a = (q31_t) (((((q63_t) a) << 32) + ((q63_t) x * y) + 0x80000000LL ) >> 32)
-
-/* SMMLSR */
-#define multSub_32x32_keep32_R(a, x, y) \
- a = (q31_t) (((((q63_t) a) << 32) - ((q63_t) x * y) + 0x80000000LL ) >> 32)
-
-/* SMMULR */
-#define mult_32x32_keep32_R(a, x, y) \
- a = (q31_t) (((q63_t) x * y + 0x80000000LL ) >> 32)
-
-/* SMMLA */
-#define multAcc_32x32_keep32(a, x, y) \
- a += (q31_t) (((q63_t) x * y) >> 32)
-
-/* SMMLS */
-#define multSub_32x32_keep32(a, x, y) \
- a -= (q31_t) (((q63_t) x * y) >> 32)
-
-/* SMMUL */
-#define mult_32x32_keep32(a, x, y) \
- a = (q31_t) (((q63_t) x * y ) >> 32)
-
-
-#define LOW_OPTIMIZATION_ENTER \
- __attribute__(( optimize("-O1") ))
-#define LOW_OPTIMIZATION_EXIT
-#define IAR_ONLY_LOW_OPTIMIZATION_ENTER
-#define IAR_ONLY_LOW_OPTIMIZATION_EXIT
-
-
-#ifdef __cplusplus
-}
-#endif
-
-
-#endif /* _RISCV_MATH_H */
-
-/**
- *
- * End of file.
- */
diff --git a/arch/risc-v/nuclei/gcc/nmsis/Library/DSP/GCC/libnmsis_dsp_rv32imac.a b/arch/risc-v/nuclei/gcc/nmsis/Library/DSP/GCC/libnmsis_dsp_rv32imac.a
deleted file mode 100644
index 3f2e5918..00000000
Binary files a/arch/risc-v/nuclei/gcc/nmsis/Library/DSP/GCC/libnmsis_dsp_rv32imac.a and /dev/null differ
diff --git a/arch/risc-v/nuclei/gcc/nmsis/Library/DSP/GCC/libnmsis_dsp_rv32imacp.a b/arch/risc-v/nuclei/gcc/nmsis/Library/DSP/GCC/libnmsis_dsp_rv32imacp.a
deleted file mode 100644
index ed8e1f08..00000000
Binary files a/arch/risc-v/nuclei/gcc/nmsis/Library/DSP/GCC/libnmsis_dsp_rv32imacp.a and /dev/null differ
diff --git a/arch/risc-v/nuclei/gcc/nmsis/Library/DSP/GCC/libnmsis_dsp_rv32imafc.a b/arch/risc-v/nuclei/gcc/nmsis/Library/DSP/GCC/libnmsis_dsp_rv32imafc.a
deleted file mode 100644
index 050b9148..00000000
Binary files a/arch/risc-v/nuclei/gcc/nmsis/Library/DSP/GCC/libnmsis_dsp_rv32imafc.a and /dev/null differ
diff --git a/arch/risc-v/nuclei/gcc/nmsis/Library/DSP/GCC/libnmsis_dsp_rv32imafcp.a b/arch/risc-v/nuclei/gcc/nmsis/Library/DSP/GCC/libnmsis_dsp_rv32imafcp.a
deleted file mode 100644
index 8adfbd24..00000000
Binary files a/arch/risc-v/nuclei/gcc/nmsis/Library/DSP/GCC/libnmsis_dsp_rv32imafcp.a and /dev/null differ
diff --git a/arch/risc-v/nuclei/gcc/nmsis/Library/DSP/GCC/libnmsis_dsp_rv32imafdc.a b/arch/risc-v/nuclei/gcc/nmsis/Library/DSP/GCC/libnmsis_dsp_rv32imafdc.a
deleted file mode 100644
index 262e9e4e..00000000
Binary files a/arch/risc-v/nuclei/gcc/nmsis/Library/DSP/GCC/libnmsis_dsp_rv32imafdc.a and /dev/null differ
diff --git a/arch/risc-v/nuclei/gcc/nmsis/Library/DSP/GCC/libnmsis_dsp_rv32imafdcp.a b/arch/risc-v/nuclei/gcc/nmsis/Library/DSP/GCC/libnmsis_dsp_rv32imafdcp.a
deleted file mode 100644
index c08ad914..00000000
Binary files a/arch/risc-v/nuclei/gcc/nmsis/Library/DSP/GCC/libnmsis_dsp_rv32imafdcp.a and /dev/null differ
diff --git a/arch/risc-v/nuclei/gcc/nmsis/Library/DSP/GCC/libnmsis_dsp_rv64imac.a b/arch/risc-v/nuclei/gcc/nmsis/Library/DSP/GCC/libnmsis_dsp_rv64imac.a
deleted file mode 100644
index 04e3709b..00000000
Binary files a/arch/risc-v/nuclei/gcc/nmsis/Library/DSP/GCC/libnmsis_dsp_rv64imac.a and /dev/null differ
diff --git a/arch/risc-v/nuclei/gcc/nmsis/Library/DSP/GCC/libnmsis_dsp_rv64imacp.a b/arch/risc-v/nuclei/gcc/nmsis/Library/DSP/GCC/libnmsis_dsp_rv64imacp.a
deleted file mode 100644
index 7f2d0706..00000000
Binary files a/arch/risc-v/nuclei/gcc/nmsis/Library/DSP/GCC/libnmsis_dsp_rv64imacp.a and /dev/null differ
diff --git a/arch/risc-v/nuclei/gcc/nmsis/Library/DSP/GCC/libnmsis_dsp_rv64imafc.a b/arch/risc-v/nuclei/gcc/nmsis/Library/DSP/GCC/libnmsis_dsp_rv64imafc.a
deleted file mode 100644
index 577a9072..00000000
Binary files a/arch/risc-v/nuclei/gcc/nmsis/Library/DSP/GCC/libnmsis_dsp_rv64imafc.a and /dev/null differ
diff --git a/arch/risc-v/nuclei/gcc/nmsis/Library/DSP/GCC/libnmsis_dsp_rv64imafcp.a b/arch/risc-v/nuclei/gcc/nmsis/Library/DSP/GCC/libnmsis_dsp_rv64imafcp.a
deleted file mode 100644
index 7a4731cf..00000000
Binary files a/arch/risc-v/nuclei/gcc/nmsis/Library/DSP/GCC/libnmsis_dsp_rv64imafcp.a and /dev/null differ
diff --git a/arch/risc-v/nuclei/gcc/nmsis/Library/DSP/GCC/libnmsis_dsp_rv64imafdc.a b/arch/risc-v/nuclei/gcc/nmsis/Library/DSP/GCC/libnmsis_dsp_rv64imafdc.a
deleted file mode 100644
index a77b5705..00000000
Binary files a/arch/risc-v/nuclei/gcc/nmsis/Library/DSP/GCC/libnmsis_dsp_rv64imafdc.a and /dev/null differ
diff --git a/arch/risc-v/nuclei/gcc/nmsis/Library/DSP/GCC/libnmsis_dsp_rv64imafdcp.a b/arch/risc-v/nuclei/gcc/nmsis/Library/DSP/GCC/libnmsis_dsp_rv64imafdcp.a
deleted file mode 100644
index d7ebb92a..00000000
Binary files a/arch/risc-v/nuclei/gcc/nmsis/Library/DSP/GCC/libnmsis_dsp_rv64imafdcp.a and /dev/null differ
diff --git a/arch/risc-v/nuclei/gcc/nmsis/Library/NN/GCC/libnmsis_nn_rv32imac.a b/arch/risc-v/nuclei/gcc/nmsis/Library/NN/GCC/libnmsis_nn_rv32imac.a
deleted file mode 100644
index 0c53f849..00000000
Binary files a/arch/risc-v/nuclei/gcc/nmsis/Library/NN/GCC/libnmsis_nn_rv32imac.a and /dev/null differ
diff --git a/arch/risc-v/nuclei/gcc/nmsis/Library/NN/GCC/libnmsis_nn_rv32imacp.a b/arch/risc-v/nuclei/gcc/nmsis/Library/NN/GCC/libnmsis_nn_rv32imacp.a
deleted file mode 100644
index 2b4bb7ac..00000000
Binary files a/arch/risc-v/nuclei/gcc/nmsis/Library/NN/GCC/libnmsis_nn_rv32imacp.a and /dev/null differ
diff --git a/arch/risc-v/nuclei/gcc/nmsis/Library/NN/GCC/libnmsis_nn_rv32imafc.a b/arch/risc-v/nuclei/gcc/nmsis/Library/NN/GCC/libnmsis_nn_rv32imafc.a
deleted file mode 100644
index f879b126..00000000
Binary files a/arch/risc-v/nuclei/gcc/nmsis/Library/NN/GCC/libnmsis_nn_rv32imafc.a and /dev/null differ
diff --git a/arch/risc-v/nuclei/gcc/nmsis/Library/NN/GCC/libnmsis_nn_rv32imafcp.a b/arch/risc-v/nuclei/gcc/nmsis/Library/NN/GCC/libnmsis_nn_rv32imafcp.a
deleted file mode 100644
index 43aff14b..00000000
Binary files a/arch/risc-v/nuclei/gcc/nmsis/Library/NN/GCC/libnmsis_nn_rv32imafcp.a and /dev/null differ
diff --git a/arch/risc-v/nuclei/gcc/nmsis/Library/NN/GCC/libnmsis_nn_rv32imafdc.a b/arch/risc-v/nuclei/gcc/nmsis/Library/NN/GCC/libnmsis_nn_rv32imafdc.a
deleted file mode 100644
index 4211799a..00000000
Binary files a/arch/risc-v/nuclei/gcc/nmsis/Library/NN/GCC/libnmsis_nn_rv32imafdc.a and /dev/null differ
diff --git a/arch/risc-v/nuclei/gcc/nmsis/Library/NN/GCC/libnmsis_nn_rv32imafdcp.a b/arch/risc-v/nuclei/gcc/nmsis/Library/NN/GCC/libnmsis_nn_rv32imafdcp.a
deleted file mode 100644
index 09793aea..00000000
Binary files a/arch/risc-v/nuclei/gcc/nmsis/Library/NN/GCC/libnmsis_nn_rv32imafdcp.a and /dev/null differ
diff --git a/arch/risc-v/nuclei/gcc/nmsis/Library/NN/GCC/libnmsis_nn_rv64imac.a b/arch/risc-v/nuclei/gcc/nmsis/Library/NN/GCC/libnmsis_nn_rv64imac.a
deleted file mode 100644
index 6868dbe2..00000000
Binary files a/arch/risc-v/nuclei/gcc/nmsis/Library/NN/GCC/libnmsis_nn_rv64imac.a and /dev/null differ
diff --git a/arch/risc-v/nuclei/gcc/nmsis/Library/NN/GCC/libnmsis_nn_rv64imacp.a b/arch/risc-v/nuclei/gcc/nmsis/Library/NN/GCC/libnmsis_nn_rv64imacp.a
deleted file mode 100644
index 42d38954..00000000
Binary files a/arch/risc-v/nuclei/gcc/nmsis/Library/NN/GCC/libnmsis_nn_rv64imacp.a and /dev/null differ
diff --git a/arch/risc-v/nuclei/gcc/nmsis/Library/NN/GCC/libnmsis_nn_rv64imafc.a b/arch/risc-v/nuclei/gcc/nmsis/Library/NN/GCC/libnmsis_nn_rv64imafc.a
deleted file mode 100644
index e79348fb..00000000
Binary files a/arch/risc-v/nuclei/gcc/nmsis/Library/NN/GCC/libnmsis_nn_rv64imafc.a and /dev/null differ
diff --git a/arch/risc-v/nuclei/gcc/nmsis/Library/NN/GCC/libnmsis_nn_rv64imafcp.a b/arch/risc-v/nuclei/gcc/nmsis/Library/NN/GCC/libnmsis_nn_rv64imafcp.a
deleted file mode 100644
index 08d9c723..00000000
Binary files a/arch/risc-v/nuclei/gcc/nmsis/Library/NN/GCC/libnmsis_nn_rv64imafcp.a and /dev/null differ
diff --git a/arch/risc-v/nuclei/gcc/nmsis/Library/NN/GCC/libnmsis_nn_rv64imafdc.a b/arch/risc-v/nuclei/gcc/nmsis/Library/NN/GCC/libnmsis_nn_rv64imafdc.a
deleted file mode 100644
index 2eb57b6e..00000000
Binary files a/arch/risc-v/nuclei/gcc/nmsis/Library/NN/GCC/libnmsis_nn_rv64imafdc.a and /dev/null differ
diff --git a/arch/risc-v/nuclei/gcc/nmsis/Library/NN/GCC/libnmsis_nn_rv64imafdcp.a b/arch/risc-v/nuclei/gcc/nmsis/Library/NN/GCC/libnmsis_nn_rv64imafdcp.a
deleted file mode 100644
index 08360fb4..00000000
Binary files a/arch/risc-v/nuclei/gcc/nmsis/Library/NN/GCC/libnmsis_nn_rv64imafdcp.a and /dev/null differ
diff --git a/arch/risc-v/nuclei/gcc/nmsis/Library/mathlib/GCC/libmathlib_rv64imafdcpv.a b/arch/risc-v/nuclei/gcc/nmsis/Library/mathlib/GCC/libmathlib_rv64imafdcpv.a
deleted file mode 100644
index c6cea696..00000000
Binary files a/arch/risc-v/nuclei/gcc/nmsis/Library/mathlib/GCC/libmathlib_rv64imafdcpv.a and /dev/null differ
diff --git a/arch/risc-v/nuclei/gcc/nmsis/NN/Include/riscv_nn_tables.h b/arch/risc-v/nuclei/gcc/nmsis/NN/Include/riscv_nn_tables.h
deleted file mode 100644
index 3b068f5f..00000000
--- a/arch/risc-v/nuclei/gcc/nmsis/NN/Include/riscv_nn_tables.h
+++ /dev/null
@@ -1,57 +0,0 @@
-/* ----------------------------------------------------------------------
- * Project: NMSIS NN Library
- * Title: riscv_nn_tables.h
- * Description: Extern declaration for NN tables
- *
- * $Date: 17. January 2018
- * $Revision: V.1.0.0
- *
- * Target Processor: RISC-V Cores
- * -------------------------------------------------------------------- */
-/*
- * Copyright (C) 2010-2018 Arm Limited or its affiliates. All rights reserved.
- * Copyright (c) 2019 Nuclei Limited. All rights reserved.
- *
- * SPDX-License-Identifier: Apache-2.0
- *
- * Licensed under the Apache License, Version 2.0 (the License); you may
- * not use this file except in compliance with the License.
- * You may obtain a copy of the License at
- *
- * www.apache.org/licenses/LICENSE-2.0
- *
- * Unless required by applicable law or agreed to in writing, software
- * distributed under the License is distributed on an AS IS BASIS, WITHOUT
- * WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
- * See the License for the specific language governing permissions and
- * limitations under the License.
- */
-
-#ifndef _RISCV_NN_TABLES_H
-#define _RISCV_NN_TABLES_H
-
-#include "riscv_math.h"
-
-/**
-* @brief tables for various activation functions
-*
-*/
-
-extern const q15_t sigmoidTable_q15[256];
-extern const q7_t sigmoidTable_q7[256];
-
-extern const q7_t tanhTable_q7[256];
-extern const q15_t tanhTable_q15[256];
-
- /**
- * @brief 2-way tables for various activation functions
- *
- * 2-way table, H table for value larger than 1/4
- * L table for value smaller than 1/4, H table for remaining
- * We have this only for the q15_t version. It does not make
- * sense to have it for q7_t type
- */
-extern const q15_t sigmoidHTable_q15[192];
-extern const q15_t sigmoidLTable_q15[128];
-
-#endif /* RISCV_NN_TABLES_H */
diff --git a/arch/risc-v/nuclei/gcc/nmsis/NN/Include/riscv_nnfunctions.h b/arch/risc-v/nuclei/gcc/nmsis/NN/Include/riscv_nnfunctions.h
deleted file mode 100644
index de7c3547..00000000
--- a/arch/risc-v/nuclei/gcc/nmsis/NN/Include/riscv_nnfunctions.h
+++ /dev/null
@@ -1,1134 +0,0 @@
-/*
- * Copyright (C) 2010-2018 Arm Limited or its affiliates. All rights reserved.
- * Copyright (c) 2019 Nuclei Limited. All rights reserved.
- *
- * SPDX-License-Identifier: Apache-2.0
- *
- * Licensed under the Apache License, Version 2.0 (the License); you may
- * not use this file except in compliance with the License.
- * You may obtain a copy of the License at
- *
- * www.apache.org/licenses/LICENSE-2.0
- *
- * Unless required by applicable law or agreed to in writing, software
- * distributed under the License is distributed on an AS IS BASIS, WITHOUT
- * WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
- * See the License for the specific language governing permissions and
- * limitations under the License.
- */
-
-/* ----------------------------------------------------------------------
- * Project: NMSIS NN Library
- * Title: riscv_nnfunctions.h
- * Description: Public header file for NMSIS NN Library
- *
- * $Date: 13. July 2018
- * $Revision: V.1.0.0
- *
- * Target Processor: RISC-V Cores
- * -------------------------------------------------------------------- */
-
-/**
- \mainpage NMSIS NN Software Library
- *
- * Introduction
- * ------------
- *
- * This user manual describes the NMSIS NN software library,
- * a collection of efficient neural network kernels developed to maximize the
- * performance and minimize the memory footprint of neural networks on Nuclei N processor cores.
- *
- * The library is divided into a number of functions each covering a specific category:
- * - Neural Network Convolution Functions
- * - Neural Network Activation Functions
- * - Fully-connected Layer Functions
- * - Neural Network Pooling Functions
- * - Softmax Functions
- * - Neural Network Support Functions
- *
- * The library has separate functions for operating on different weight and activation data
- * types including 8-bit integers (q7_t) and 16-bit integers (q15_t). The description of the
- * kernels are included in the function description. The implementation details are also
- * described in this paper [1].
- *
- * \note Please refer to [NMSIS-NN](../../../nn/index.html)
- *
- * Block Diagram
- * --------
- * \image html NMSIS-NN-OVERVIEW.PNG
- *
- * Examples
- * --------
- *
- * The library ships with a number of examples which demonstrate how to use the library functions.
- *
- * Pre-processor Macros
- * ------------
- *
- * Each library project have differant pre-processor macros.
- *
- * - RISCV_MATH_DSP:
- *
- * Define macro RISCV_MATH_DSP, If the silicon supports DSP instructions.
- *
- * - RISCV_NN_TRUNCATE:
- *
- * Define macro RISCV_NN_TRUNCATE to use floor instead of round-to-the-nearest-int for the computation.
- *
- *
- * [1] CMSIS-NN: Efficient Neural Network Kernels for Arm Cortex-M CPUs https://arxiv.org/abs/1801.06601
- */
-
-/**
- * @defgroup groupNN Neural Network Functions
- * These functions perform basic operations for neural network layers.
- */
-
-#ifndef _RISCV_NNFUNCTIONS_H
-#define _RISCV_NNFUNCTIONS_H
-
-#include "riscv_nnsupportfunctions.h"
-#include "riscv_nn_tables.h"
-
-#define USE_INTRINSIC
-
-//#define RISCV_NN_TRUNCATE /* This config the rounding model to floor or round to the nearest int */
-
-#ifdef __cplusplus
-extern "C"
-{
-#endif
-
-/**
- * @defgroup NNConv Neural Network Convolution Functions
- *
- * Perform convolution layer
- *
- * The convolution is implemented in 2 steps: im2col and GEMM
- *
- * im2col is a process of converting each patch of image data into
- * a column. After im2col, the convolution is computed as matrix-matrix
- * multiplication.
- *
- * To reduce the memory footprint, the im2col is performed partially.
- * Each iteration, only a few column (i.e., patches) are generated and
- * computed with GEMM kernels similar to NMSIS-DSP riscv_mat_mult functions.
- *
- */
-
- /**
- * @brief Basic Q7 convolution function
- * @param[in] Im_in pointer to input tensor
- * @param[in] dim_im_in input tensor dimension
- * @param[in] ch_im_in number of input tensor channels
- * @param[in] wt pointer to kernel weights
- * @param[in] ch_im_out number of filters, i.e., output tensor channels
- * @param[in] dim_kernel filter kernel size
- * @param[in] padding padding sizes
- * @param[in] stride convolution stride
- * @param[in] bias pointer to bias
- * @param[in] bias_shift amount of left-shift for bias
- * @param[in] out_shift amount of right-shift for output
- * @param[in,out] Im_out pointer to output tensor
- * @param[in] dim_im_out output tensor dimension
- * @param[in,out] bufferA pointer to buffer space for input
- * @param[in,out] bufferB pointer to buffer space for output
- * @return The function returns RISCV_MATH_SUCCESS
- *
- */
-
- riscv_status riscv_convolve_HWC_q7_basic(const q7_t * Im_in,
- const uint16_t dim_im_in,
- const uint16_t ch_im_in,
- const q7_t * wt,
- const uint16_t ch_im_out,
- const uint16_t dim_kernel,
- const uint16_t padding,
- const uint16_t stride,
- const q7_t * bias,
- const uint16_t bias_shift,
- const uint16_t out_shift,
- q7_t * Im_out,
- const uint16_t dim_im_out,
- q15_t * bufferA,
- q7_t * bufferB);
-
- /**
- * @brief Basic Q7 convolution function (non-sqaure shape)
- * @param[in] Im_in pointer to input tensor
- * @param[in] dim_im_in_x input tensor dimension x
- * @param[in] dim_im_in_y input tensor dimension y
- * @param[in] ch_im_in number of input tensor channels
- * @param[in] wt pointer to kernel weights
- * @param[in] ch_im_out number of filters, i.e., output tensor channels
- * @param[in] dim_kernel_x filter kernel size x
- * @param[in] dim_kernel_y filter kernel size y
- * @param[in] padding_x padding size x
- * @param[in] padding_y padding size y
- * @param[in] stride_x convolution stride x
- * @param[in] stride_y convolution stride y
- * @param[in] bias pointer to bias
- * @param[in] bias_shift amount of left-shift for bias
- * @param[in] out_shift amount of right-shift for output
- * @param[in,out] Im_out pointer to output tensor
- * @param[in] dim_im_out_x output tensor dimension x
- * @param[in] dim_im_out_y output tensor dimension y
- * @param[in,out] bufferA pointer to buffer space for input
- * @param[in,out] bufferB pointer to buffer space for output
- * @return The function returns RISCV_MATH_SUCCESS
- */
-
- riscv_status riscv_convolve_HWC_q7_basic_nonsquare(const q7_t * Im_in,
- const uint16_t dim_im_in_x,
- const uint16_t dim_im_in_y,
- const uint16_t ch_im_in,
- const q7_t * wt,
- const uint16_t ch_im_out,
- const uint16_t dim_kernel_x,
- const uint16_t dim_kernel_y,
- const uint16_t padding_x,
- const uint16_t padding_y,
- const uint16_t stride_x,
- const uint16_t stride_y,
- const q7_t * bias,
- const uint16_t bias_shift,
- const uint16_t out_shift,
- q7_t * Im_out,
- const uint16_t dim_im_out_x,
- const uint16_t dim_im_out_y,
- q15_t * bufferA,
- q7_t * bufferB);
-
- /**
- * @brief Basic Q15 convolution function
- * @param[in] Im_in pointer to input tensor
- * @param[in] dim_im_in input tensor dimension
- * @param[in] ch_im_in number of input tensor channels
- * @param[in] wt pointer to kernel weights
- * @param[in] ch_im_out number of filters, i.e., output tensor channels
- * @param[in] dim_kernel filter kernel size
- * @param[in] padding padding sizes
- * @param[in] stride convolution stride
- * @param[in] bias pointer to bias
- * @param[in] bias_shift amount of left-shift for bias
- * @param[in] out_shift amount of right-shift for output
- * @param[in,out] Im_out pointer to output tensor
- * @param[in] dim_im_out output tensor dimension
- * @param[in,out] bufferA pointer to buffer space for input
- * @param[in,out] bufferB pointer to buffer space for output
- * @return The function returns RISCV_MATH_SUCCESS
- *
- */
-
- riscv_status riscv_convolve_HWC_q15_basic(const q15_t * Im_in,
- const uint16_t dim_im_in,
- const uint16_t ch_im_in,
- const q15_t * wt,
- const uint16_t ch_im_out,
- const uint16_t dim_kernel,
- const uint16_t padding,
- const uint16_t stride,
- const q15_t * bias,
- const uint16_t bias_shift,
- const uint16_t out_shift,
- q15_t * Im_out,
- const uint16_t dim_im_out,
- q15_t * bufferA,
- q7_t * bufferB);
-
- /**
- * @brief Fast Q7 convolution function
- * @param[in] Im_in pointer to input tensor
- * @param[in] dim_im_in input tensor dimension
- * @param[in] ch_im_in number of input tensor channels
- * @param[in] wt pointer to kernel weights
- * @param[in] ch_im_out number of filters, i.e., output tensor channels
- * @param[in] dim_kernel filter kernel size
- * @param[in] padding padding sizes
- * @param[in] stride convolution stride
- * @param[in] bias pointer to bias
- * @param[in] bias_shift amount of left-shift for bias
- * @param[in] out_shift amount of right-shift for output
- * @param[in,out] Im_out pointer to output tensor
- * @param[in] dim_im_out output tensor dimension
- * @param[in,out] bufferA pointer to buffer space for input
- * @param[in,out] bufferB pointer to buffer space for output
- * @return The function returns either
- * RISCV_MATH_SIZE_MISMATCH or RISCV_MATH_SUCCESS based on the outcome of size checking.
- *
- * This function is the version with full list of optimization tricks, but with
- * some constraints:
- * ch_im_in is multiple of 4
- * ch_im_out is multiple of 2
- */
-
- riscv_status riscv_convolve_HWC_q7_fast(const q7_t * Im_in,
- const uint16_t dim_im_in,
- const uint16_t ch_im_in,
- const q7_t * wt,
- const uint16_t ch_im_out,
- const uint16_t dim_kernel,
- const uint16_t padding,
- const uint16_t stride,
- const q7_t * bias,
- const uint16_t bias_shift,
- const uint16_t out_shift,
- q7_t * Im_out,
- const uint16_t dim_im_out,
- q15_t * bufferA,
- q7_t * bufferB);
-
- /**
- * @brief Fast Q7 convolution function (non-sqaure shape)
- * @param[in] Im_in pointer to input tensor
- * @param[in] dim_im_in_x input tensor dimension x
- * @param[in] dim_im_in_y input tensor dimension y
- * @param[in] ch_im_in number of input tensor channels
- * @param[in] wt pointer to kernel weights
- * @param[in] ch_im_out number of filters, i.e., output tensor channels
- * @param[in] dim_kernel_x filter kernel size x
- * @param[in] dim_kernel_y filter kernel size y
- * @param[in] padding_x padding size x
- * @param[in] padding_y padding size y
- * @param[in] stride_x convolution stride x
- * @param[in] stride_y convolution stride y
- * @param[in] bias pointer to bias
- * @param[in] bias_shift amount of left-shift for bias
- * @param[in] out_shift amount of right-shift for output
- * @param[in,out] Im_out pointer to output tensor
- * @param[in] dim_im_out_x output tensor dimension x
- * @param[in] dim_im_out_y output tensor dimension y
- * @param[in,out] bufferA pointer to buffer space for input
- * @param[in,out] bufferB pointer to buffer space for output
- * @return The function returns either
- * RISCV_MATH_SIZE_MISMATCH or RISCV_MATH_SUCCESS based on the outcome of size checking.
- *
- * This function is the version with full list of optimization tricks, but with
- * some constraints:
- * ch_im_in is multiple of 4
- * ch_im_out is multiple of 2
- */
-
- riscv_status riscv_convolve_HWC_q7_fast_nonsquare(const q7_t * Im_in,
- const uint16_t dim_im_in_x,
- const uint16_t dim_im_in_y,
- const uint16_t ch_im_in,
- const q7_t * wt,
- const uint16_t ch_im_out,
- const uint16_t dim_kernel_x,
- const uint16_t dim_kernel_y,
- const uint16_t padding_x,
- const uint16_t padding_y,
- const uint16_t stride_x,
- const uint16_t stride_y,
- const q7_t * bias,
- const uint16_t bias_shift,
- const uint16_t out_shift,
- q7_t * Im_out,
- const uint16_t dim_im_out_x,
- const uint16_t dim_im_out_y,
- q15_t * bufferA,
- q7_t * bufferB);
-
- /**
- * @brief Fast Q7 version of 1x1 convolution (non-sqaure shape)
- * @param[in] Im_in pointer to input tensor
- * @param[in] dim_im_in_x input tensor dimension x
- * @param[in] dim_im_in_y input tensor dimension y
- * @param[in] ch_im_in number of input tensor channels
- * @param[in] wt pointer to kernel weights
- * @param[in] ch_im_out number of filters, i.e., output tensor channels
- * @param[in] dim_kernel_x filter kernel size x
- * @param[in] dim_kernel_y filter kernel size y
- * @param[in] padding_x padding size x
- * @param[in] padding_y padding size y
- * @param[in] stride_x convolution stride x
- * @param[in] stride_y convolution stride y
- * @param[in] bias pointer to bias
- * @param[in] bias_shift amount of left-shift for bias
- * @param[in] out_shift amount of right-shift for output
- * @param[in,out] Im_out pointer to output tensor
- * @param[in] dim_im_out_x output tensor dimension x
- * @param[in] dim_im_out_y output tensor dimension y
- * @param[in,out] bufferA pointer to buffer space for input
- * @param[in,out] bufferB pointer to buffer space for output
- * @return The function returns either
- * RISCV_MATH_SIZE_MISMATCH or RISCV_MATH_SUCCESS based on the outcome of size checking.
- *
- * This function implement convolution with 1x1 kernel size (i.e., dim_kernel_x=1
- * and dim_kernel_y=1). It can be used for
- * second half of MobileNets after depthwise separable convolution.
- *
- * This function is the version with full list of optimization tricks, but with
- * some constraints:
- * ch_im_in is multiple of 4
- * ch_im_out is multiple of 2
- */
- riscv_status riscv_convolve_1x1_HWC_q7_fast_nonsquare(const q7_t * Im_in,
- const uint16_t dim_im_in_x,
- const uint16_t dim_im_in_y,
- const uint16_t ch_im_in,
- const q7_t * wt,
- const uint16_t ch_im_out,
- const uint16_t dim_kernel_x,
- const uint16_t dim_kernel_y,
- const uint16_t padding_x,
- const uint16_t padding_y,
- const uint16_t stride_x,
- const uint16_t stride_y,
- const q7_t * bias,
- const uint16_t bias_shift,
- const uint16_t out_shift,
- q7_t * Im_out,
- const uint16_t dim_im_out_x,
- const uint16_t dim_im_out_y,
- q15_t * bufferA,
- q7_t * bufferB);
-
- /**
- * @brief Q7 version of convolution for RGB image
- * @param[in] Im_in pointer to input tensor
- * @param[in] dim_im_in input tensor dimension
- * @param[in] ch_im_in number of input tensor channels
- * @param[in] wt pointer to kernel weights
- * @param[in] ch_im_out number of filters, i.e., output tensor channels
- * @param[in] dim_kernel filter kernel size
- * @param[in] padding padding sizes
- * @param[in] stride convolution stride
- * @param[in] bias pointer to bias
- * @param[in] bias_shift amount of left-shift for bias
- * @param[in] out_shift amount of right-shift for output
- * @param[in,out] Im_out pointer to output tensor
- * @param[in] dim_im_out output tensor dimension
- * @param[in,out] bufferA pointer to buffer space for input
- * @param[in,out] bufferB pointer to buffer space for output
- * @return The function returns either
- * RISCV_MATH_SIZE_MISMATCH or RISCV_MATH_SUCCESS based on the outcome of size checking.
- *
- * This kernel is written exclusively for convolution with ch_im_in
- * equals 3. This applies on the first layer of CNNs which has input
- * image with RGB format.
- */
-
- riscv_status riscv_convolve_HWC_q7_RGB(const q7_t * Im_in,
- const uint16_t dim_im_in,
- const uint16_t ch_im_in,
- const q7_t * wt,
- const uint16_t ch_im_out,
- const uint16_t dim_kernel,
- const uint16_t padding,
- const uint16_t stride,
- const q7_t * bias,
- const uint16_t bias_shift,
- const uint16_t out_shift,
- q7_t * Im_out,
- const uint16_t dim_im_out,
- q15_t * bufferA,
- q7_t * bufferB);
-
- /**
- * @brief Fast Q15 convolution function
- * @param[in] Im_in pointer to input tensor
- * @param[in] dim_im_in input tensor dimension
- * @param[in] ch_im_in number of input tensor channels
- * @param[in] wt pointer to kernel weights
- * @param[in] ch_im_out number of filters, i.e., output tensor channels
- * @param[in] dim_kernel filter kernel size
- * @param[in] padding padding sizes
- * @param[in] stride convolution stride
- * @param[in] bias pointer to bias
- * @param[in] bias_shift amount of left-shift for bias
- * @param[in] out_shift amount of right-shift for output
- * @param[in,out] Im_out pointer to output tensor
- * @param[in] dim_im_out output tensor dimension
- * @param[in,out] bufferA pointer to buffer space for input
- * @param[in,out] bufferB pointer to buffer space for output
- * @return The function returns either
- * RISCV_MATH_SIZE_MISMATCH or RISCV_MATH_SUCCESS based on the outcome of size checking.
- *
- * This function is the version with full list of optimization tricks, but with
- * some constraints:
- * ch_im_in is multiple of 2
- * ch_im_out is multiple of 2
- */
-
- riscv_status riscv_convolve_HWC_q15_fast(const q15_t * Im_in,
- const uint16_t dim_im_in,
- const uint16_t ch_im_in,
- const q15_t * wt,
- const uint16_t ch_im_out,
- const uint16_t dim_kernel,
- const uint16_t padding,
- const uint16_t stride,
- const q15_t * bias,
- const uint16_t bias_shift,
- const uint16_t out_shift,
- q15_t * Im_out,
- const uint16_t dim_im_out,
- q15_t * bufferA,
- q7_t * bufferB);
-
- /**
- * @brief Fast Q15 convolution function (non-sqaure shape)
- * @param[in] Im_in pointer to input tensor
- * @param[in] dim_im_in_x input tensor dimension x
- * @param[in] dim_im_in_y input tensor dimension y
- * @param[in] ch_im_in number of input tensor channels
- * @param[in] wt pointer to kernel weights
- * @param[in] ch_im_out number of filters, i.e., output tensor channels
- * @param[in] dim_kernel_x filter kernel size x
- * @param[in] dim_kernel_y filter kernel size y
- * @param[in] padding_x padding size x
- * @param[in] padding_y padding size y
- * @param[in] stride_x convolution stride x
- * @param[in] stride_y convolution stride y
- * @param[in] bias pointer to bias
- * @param[in] bias_shift amount of left-shift for bias
- * @param[in] out_shift amount of right-shift for output
- * @param[in,out] Im_out pointer to output tensor
- * @param[in] dim_im_out_x output tensor dimension x
- * @param[in] dim_im_out_y output tensor dimension y
- * @param[in,out] bufferA pointer to buffer space for input
- * @param[in,out] bufferB pointer to buffer space for output
- * @return The function returns either
- * RISCV_MATH_SIZE_MISMATCH or RISCV_MATH_SUCCESS based on the outcome of size checking.
- *
- * @details
- *
- * Buffer size:
- *
- * bufferA size: 2*ch_im_in*dim_kernel*dim_kernel
- *
- * bufferB size: 0
- *
- * Input dimension constraints:
- *
- * ch_im_in is multiple of 2
- *
- * ch_im_out is multiple of 2
- *
- */
-
- riscv_status
- riscv_convolve_HWC_q15_fast_nonsquare(const q15_t * Im_in,
- const uint16_t dim_im_in_x,
- const uint16_t dim_im_in_y,
- const uint16_t ch_im_in,
- const q15_t * wt,
- const uint16_t ch_im_out,
- const uint16_t dim_kernel_x,
- const uint16_t dim_kernel_y,
- const uint16_t padding_x,
- const uint16_t padding_y,
- const uint16_t stride_x,
- const uint16_t stride_y,
- const q15_t * bias,
- const uint16_t bias_shift,
- const uint16_t out_shift,
- q15_t * Im_out,
- const uint16_t dim_im_out_x,
- const uint16_t dim_im_out_y,
- q15_t * bufferA,
- q7_t * bufferB);
-
- /**
- * @brief Q7 depthwise separable convolution function
- * @param[in] Im_in pointer to input tensor
- * @param[in] dim_im_in input tensor dimension
- * @param[in] ch_im_in number of input tensor channels
- * @param[in] wt pointer to kernel weights
- * @param[in] ch_im_out number of filters, i.e., output tensor channels
- * @param[in] dim_kernel filter kernel size
- * @param[in] padding padding sizes
- * @param[in] stride convolution stride
- * @param[in] bias pointer to bias
- * @param[in] bias_shift amount of left-shift for bias
- * @param[in] out_shift amount of right-shift for output
- * @param[in,out] Im_out pointer to output tensor
- * @param[in] dim_im_out output tensor dimension
- * @param[in,out] bufferA pointer to buffer space for input
- * @param[in,out] bufferB pointer to buffer space for output
- * @return The function returns either
- * RISCV_MATH_SIZE_MISMATCH or RISCV_MATH_SUCCESS based on the outcome of size checking.
- *
- * This function is the version with full list of optimization tricks, but with
- * some constraints:
- * ch_im_in is multiple of 2
- * ch_im_out is multiple of 2
- */
-
- riscv_status riscv_depthwise_separable_conv_HWC_q7(const q7_t * Im_in,
- const uint16_t dim_im_in,
- const uint16_t ch_im_in,
- const q7_t * wt,
- const uint16_t ch_im_out,
- const uint16_t dim_kernel,
- const uint16_t padding,
- const uint16_t stride,
- const q7_t * bias,
- const uint16_t bias_shift,
- const uint16_t out_shift,
- q7_t * Im_out,
- const uint16_t dim_im_out,
- q15_t * bufferA,
- q7_t * bufferB);
-
- /**
- * @brief Q7 depthwise separable convolution function (non-square shape)
- * @param[in] Im_in pointer to input tensor
- * @param[in] dim_im_in_x input tensor dimension x
- * @param[in] dim_im_in_y input tensor dimension y
- * @param[in] ch_im_in number of input tensor channels
- * @param[in] wt pointer to kernel weights
- * @param[in] ch_im_out number of filters, i.e., output tensor channels
- * @param[in] dim_kernel_x filter kernel size x
- * @param[in] dim_kernel_y filter kernel size y
- * @param[in] padding_x padding sizes x
- * @param[in] padding_y padding sizes y
- * @param[in] stride_x convolution stride x
- * @param[in] stride_y convolution stride y
- * @param[in] bias pointer to bias
- * @param[in] bias_shift amount of left-shift for bias
- * @param[in] out_shift amount of right-shift for output
- * @param[in,out] Im_out pointer to output tensor
- * @param[in] dim_im_out_x output tensor dimension x
- * @param[in] dim_im_out_y output tensor dimension y
- * @param[in,out] bufferA pointer to buffer space for input
- * @param[in,out] bufferB pointer to buffer space for output
- * @return The function returns either
- * RISCV_MATH_SIZE_MISMATCH or RISCV_MATH_SUCCESS based on the outcome of size checking.
- *
- * This function is the version with full list of optimization tricks, but with
- * some constraints:
- * ch_im_in is multiple of 2
- * ch_im_out is multiple of 2
- */
- riscv_status riscv_depthwise_separable_conv_HWC_q7_nonsquare(const q7_t * Im_in,
- const uint16_t dim_im_in_x,
- const uint16_t dim_im_in_y,
- const uint16_t ch_im_in,
- const q7_t * wt,
- const uint16_t ch_im_out,
- const uint16_t dim_kernel_x,
- const uint16_t dim_kernel_y,
- const uint16_t padding_x,
- const uint16_t padding_y,
- const uint16_t stride_x,
- const uint16_t stride_y,
- const q7_t * bias,
- const uint16_t bias_shift,
- const uint16_t out_shift,
- q7_t * Im_out,
- const uint16_t dim_im_out_x,
- const uint16_t dim_im_out_y,
- q15_t * bufferA,
- q7_t * bufferB);
-
-
-/**
- * @defgroup FC Fully-connected Layer Functions
- *
- * Perform fully-connected layer
- *
- * Fully-connected layer is basically a matrix-vector multiplication
- * with bias. The matrix is the weights and the input/output vectors
- * are the activation values. Supported {weight, activation} precisions
- * include {8-bit, 8-bit}, {16-bit, 16-bit}, and {8-bit, 16-bit}.
- *
- * Here we have two types of kernel functions. The basic function
- * implements the function using regular GEMV approach. The opt functions
- * operates with weights in interleaved formats.
- *
- */
-
- /**
- * @brief Q7 basic fully-connected layer function
- * @param[in] pV pointer to input vector
- * @param[in] pM pointer to matrix weights
- * @param[in] dim_vec length of the vector
- * @param[in] num_of_rows number of rows in weight matrix
- * @param[in] bias_shift amount of left-shift for bias
- * @param[in] out_shift amount of right-shift for output
- * @param[in] bias pointer to bias
- * @param[in,out] pOut pointer to output vector
- * @param[in,out] vec_buffer pointer to buffer space for input
- * @return The function returns RISCV_MATH_SUCCESS
- *
- */
-
- riscv_status riscv_fully_connected_q7(const q7_t * pV,
- const q7_t * pM,
- const uint16_t dim_vec,
- const uint16_t num_of_rows,
- const uint16_t bias_shift,
- const uint16_t out_shift,
- const q7_t * bias,
- q7_t * pOut,
- q15_t * vec_buffer);
-
- /**
- * @brief S8 basic fully-connected layer function for TF Lite
- * @param[in] pInput pointer to pInput vector
- * @param[in] pWeight pointer to matrix weights
- * @param[in] col_dim dimension of the input vector
- * @param[in] row_dim dimension of the output vector
- * @param[in] nb_batches number of batches
- * @param[in] input_offset
- * @param[in] filter_offset
- * @param[in] out_mult requantization parameter
- * @param[in] out_shift requantization parameter
- * @param[in] output_offset
- * @param[in] pBias pointer to bias
- * @param[out] pOut pointer to output vector
- * @param[in] output_activation_min for clamping
- * @param[in] output_activation_max for clamping
- * @param[in,out] vec_buffer pointer to buffer space for pInput
- * @return The function returns RISCV_MATH_SUCCESS
- *
- * @details
- *
- * Buffer size:
- *
- * vec_buffer size: col_dim of word16.
- *
- * This basic function is designed to work with regular pWeight
- * matrix without interleaving.
- *
- */
- riscv_status
- riscv_fully_connected_s8(const int8_t *pInput,
- const int8_t *weight,
- const uint16_t input_length,
- const uint16_t num_rows,
- const uint16_t nb_batches,
- const int32_t input_offset,
- const int32_t filter_offset,
- const int32_t out_mult,
- const int32_t out_shift,
- const int32_t output_offset,
- const int8_t *bias,
- int8_t *pOut,
- const int32_t output_activation_min,
- const int32_t output_activation_max,
- q15_t *vec_buffer) ;
-
- /**
- * @brief Q7 opt fully-connected layer function
- * @param[in] pV pointer to input vector
- * @param[in] pM pointer to matrix weights
- * @param[in] dim_vec length of the vector
- * @param[in] num_of_rows number of rows in weight matrix
- * @param[in] bias_shift amount of left-shift for bias
- * @param[in] out_shift amount of right-shift for output
- * @param[in] bias pointer to bias
- * @param[in,out] pOut pointer to output vector
- * @param[in,out] vec_buffer pointer to buffer space for input
- * @return The function returns RISCV_MATH_SUCCESS
- *
- */
-
- riscv_status riscv_fully_connected_q7_opt(const q7_t * pV,
- const q7_t * pM,
- const uint16_t dim_vec,
- const uint16_t num_of_rows,
- const uint16_t bias_shift,
- const uint16_t out_shift,
- const q7_t * bias,
- q7_t * pOut,
- q15_t * vec_buffer);
-
- /**
- * @brief Q15 basic fully-connected layer function
- * @param[in] pV pointer to input vector
- * @param[in] pM pointer to matrix weights
- * @param[in] dim_vec length of the vector
- * @param[in] num_of_rows number of rows in weight matrix
- * @param[in] bias_shift amount of left-shift for bias
- * @param[in] out_shift amount of right-shift for output
- * @param[in] bias pointer to bias
- * @param[in,out] pOut pointer to output vector
- * @param[in,out] vec_buffer pointer to buffer space for input
- * @return The function returns RISCV_MATH_SUCCESS
- *
- */
-
- riscv_status riscv_fully_connected_q15(const q15_t * pV,
- const q15_t * pM,
- const uint16_t dim_vec,
- const uint16_t num_of_rows,
- const uint16_t bias_shift,
- const uint16_t out_shift,
- const q15_t * bias,
- q15_t * pOut,
- q15_t * vec_buffer);
-
- /**
- * @brief Q15 opt fully-connected layer function
- * @param[in] pV pointer to input vector
- * @param[in] pM pointer to matrix weights
- * @param[in] dim_vec length of the vector
- * @param[in] num_of_rows number of rows in weight matrix
- * @param[in] bias_shift amount of left-shift for bias
- * @param[in] out_shift amount of right-shift for output
- * @param[in] bias pointer to bias
- * @param[in,out] pOut pointer to output vector
- * @param[in,out] vec_buffer pointer to buffer space for input
- * @return The function returns RISCV_MATH_SUCCESS
- *
- */
-
- riscv_status riscv_fully_connected_q15_opt(const q15_t * pV,
- const q15_t * pM,
- const uint16_t dim_vec,
- const uint16_t num_of_rows,
- const uint16_t bias_shift,
- const uint16_t out_shift,
- const q15_t * bias,
- q15_t * pOut,
- q15_t * vec_buffer);
-
- /**
- * @brief Mixed Q15-Q7 fully-connected layer function
- * @param[in] pV pointer to input vector
- * @param[in] pM pointer to matrix weights
- * @param[in] dim_vec length of the vector
- * @param[in] num_of_rows number of rows in weight matrix
- * @param[in] bias_shift amount of left-shift for bias
- * @param[in] out_shift amount of right-shift for output
- * @param[in] bias pointer to bias
- * @param[in,out] pOut pointer to output vector
- * @param[in,out] vec_buffer pointer to buffer space for input
- * @return The function returns RISCV_MATH_SUCCESS
- *
- */
-
- riscv_status riscv_fully_connected_mat_q7_vec_q15(const q15_t * pV,
- const q7_t * pM,
- const uint16_t dim_vec,
- const uint16_t num_of_rows,
- const uint16_t bias_shift,
- const uint16_t out_shift,
- const q7_t * bias,
- q15_t * pOut,
- q15_t * vec_buffer);
-
- /**
- * @brief Mixed Q15-Q7 opt fully-connected layer function
- * @param[in] pV pointer to input vector
- * @param[in] pM pointer to matrix weights
- * @param[in] dim_vec length of the vector
- * @param[in] num_of_rows number of rows in weight matrix
- * @param[in] bias_shift amount of left-shift for bias
- * @param[in] out_shift amount of right-shift for output
- * @param[in] bias pointer to bias
- * @param[in,out] pOut pointer to output vector
- * @param[in,out] vec_buffer pointer to buffer space for input
- * @return The function returns RISCV_MATH_SUCCESS
- *
- */
-
- riscv_status riscv_fully_connected_mat_q7_vec_q15_opt(const q15_t * pV,
- const q7_t * pM,
- const uint16_t dim_vec,
- const uint16_t num_of_rows,
- const uint16_t bias_shift,
- const uint16_t out_shift,
- const q7_t * bias,
- q15_t * pOut,
- q15_t * vec_buffer);
-
-/**
- * @brief Matrix-Multiplication Kernels for Convolution
- *
- * These functions are used within convolution layer functions for
- * matrix multiplication.
- *
- * The implementation is similar to NMSIS-DSP riscv_mat_mult functions
- * with one Q7 and one Q15 operands. The Q15 operand is the im2col
- * output which is always with 2 columns.
- *
- */
-
- /**
- * @brief Matrix-multiplication function for convolution
- * @param[in] pA pointer to operand A
- * @param[in] pInBuffer pointer to operand B, always conssists of 2 vectors
- * @param[in] ch_im_out numRow of A
- * @param[in] numCol_A numCol of A
- * @param[in] bias_shift amount of left-shift for bias
- * @param[in] out_shift amount of right-shift for output
- * @param[in] bias the bias
- * @param[in,out] pOut pointer to output
- * @return The function returns the incremented output pointer
- */
-
- q7_t *riscv_nn_mat_mult_kernel_q7_q15(const q7_t * pA,
- const q15_t * pInBuffer,
- const uint16_t ch_im_out,
- const uint16_t numCol_A,
- const uint16_t bias_shift,
- const uint16_t out_shift,
- const q7_t * bias,
- q7_t * pOut);
-
- q7_t *riscv_nn_mat_mult_kernel_q7(const q7_t * pA,
- const q7_t * pInBuffer,
- const uint16_t ch_im_out,
- const uint16_t numCol_A,
- const uint16_t bias_shift,
- const uint16_t out_shift,
- const q7_t * bias,
- q7_t * pOut);
-
- /**
- * @brief Matrix-multiplication function for convolution with reordered columns
- * @param[in] pA pointer to operand A
- * @param[in] pInBuffer pointer to operand B, always conssists of 2 vectors
- * @param[in] ch_im_out numRow of A
- * @param[in] numCol_A numCol of A
- * @param[in] bias_shift amount of left-shift for bias
- * @param[in] out_shift amount of right-shift for output
- * @param[in] bias the bias
- * @param[in,out] pOut pointer to output
- * @return The function returns the incremented output pointer
- */
-
- q7_t *riscv_nn_mat_mult_kernel_q7_q15_reordered(const q7_t * pA,
- const q15_t * pInBuffer,
- const uint16_t ch_im_out,
- const uint16_t numCol_A,
- const uint16_t bias_shift,
- const uint16_t out_shift,
- const q7_t * bias,
- q7_t * pOut);
-
- q7_t *riscv_nn_mat_mult_kernel_q7_reordered(const q7_t * pA,
- const q7_t * pInBuffer,
- const uint16_t ch_im_out,
- const uint16_t numCol_A,
- const uint16_t bias_shift,
- const uint16_t out_shift,
- const q7_t * bias,
- q7_t * pOut);
-
-#ifdef __cplusplus
-}
-#endif
-
-/*
- * Other functions
- * These layers are typically not timing critical
- * Basic implementation is supported here
- */
-
-#ifdef __cplusplus
-extern "C"
-{
-#endif
-
-/**
- * @defgroup Acti Neural Network Activation Functions
- *
- * Perform activation layers, including ReLU (Rectified Linear Unit),
- * sigmoid and tanh
- *
- */
-
- /**
- * @brief Q7 RELU function
- * @param[in,out] data pointer to input
- * @param[in] size number of elements
- * @return none.
- */
-
- void riscv_relu_q7(q7_t * data, uint16_t size);
-
- /**
- * @brief Q15 RELU function
- * @param[in,out] data pointer to input
- * @param[in] size number of elements
- * @return none.
- */
-
- void riscv_relu_q15(q15_t * data, uint16_t size);
-
- /**
- * @brief Q7 neural network activation function using direct table look-up
- * @param[in,out] data pointer to input
- * @param[in] size number of elements
- * @param[in] int_width bit-width of the integer part, assume to be smaller than 3
- * @param[in] type type of activation functions
- * @return none.
- */
-
- void riscv_nn_activations_direct_q7(q7_t * data, uint16_t size, uint16_t int_width,
- riscv_nn_activation_type type);
-
- /**
- * @brief Q15 neural network activation function using direct table look-up
- * @param[in,out] data pointer to input
- * @param[in] size number of elements
- * @param[in] int_width bit-width of the integer part, assume to be smaller than 3
- * @param[in] type type of activation functions
- * @return none.
- */
-
- void riscv_nn_activations_direct_q15(q15_t * data, uint16_t size, uint16_t int_width,
- riscv_nn_activation_type type);
-
-/**
- * @defgroup Pooling Neural Network Pooling Functions
- *
- * Perform pooling functions, including max pooling and average pooling
- *
- */
-
- /**
- * @brief Q7 max pooling function
- * @param[in] Im_in pointer to input tensor
- * @param[in] dim_im_in input tensor dimension
- * @param[in] ch_im_in number of input tensor channels
- * @param[in] dim_kernel filter kernel size
- * @param[in] padding padding sizes
- * @param[in] stride convolution stride
- * @param[in] dim_im_out output tensor dimension
- * @param[in,out] bufferA pointer to buffer space for input
- * @param[in,out] Im_out pointer to output tensor
- * @return none.
- *
- */
-
- void riscv_maxpool_q7_HWC(q7_t * Im_in,
- const uint16_t dim_im_in,
- const uint16_t ch_im_in,
- const uint16_t dim_kernel,
- const uint16_t padding,
- const uint16_t stride,
- const uint16_t dim_im_out,
- q7_t * bufferA,
- q7_t * Im_out);
-
- /**
- * @brief Q7 average pooling function
- * @param[in] Im_in pointer to input tensor
- * @param[in] dim_im_in input tensor dimension
- * @param[in] ch_im_in number of input tensor channels
- * @param[in] dim_kernel filter kernel size
- * @param[in] padding padding sizes
- * @param[in] stride convolution stride
- * @param[in] dim_im_out output tensor dimension
- * @param[in,out] bufferA pointer to buffer space for input
- * @param[in,out] Im_out pointer to output tensor
- * @return none.
- *
- */
-
- void riscv_avepool_q7_HWC(q7_t * Im_in,
- const uint16_t dim_im_in,
- const uint16_t ch_im_in,
- const uint16_t dim_kernel,
- const uint16_t padding,
- const uint16_t stride,
- const uint16_t dim_im_out,
- q7_t * bufferA,
- q7_t * Im_out);
-
-/**
- * @defgroup Softmax Softmax Functions
- *
- * EXP(2) based softmax function
- *
- */
-
- /**
- * @brief Q7 softmax function
- * @param[in] vec_in pointer to input vector
- * @param[in] dim_vec input vector dimension
- * @param[out] p_out pointer to output vector
- * @return none.
- *
- */
-
- void riscv_softmax_q7(const q7_t * vec_in, const uint16_t dim_vec, q7_t * p_out);
-
- /**
- * @brief Q15 softmax function
- * @param[in] vec_in pointer to input vector
- * @param[in] dim_vec input vector dimension
- * @param[out] p_out pointer to output vector
- * @return none.
- *
- */
-
- void riscv_softmax_q15(const q15_t * vec_in, const uint16_t dim_vec, q15_t * p_out);
-
- /**
- * @brief uint8 depthwise convolution function with asymmetric quantization for even number of channel multiplier
- * and input channels. Unless specified otherwise, arguments are mandatory.
- *
- * @param[in] input Pointer to input tensor
- * @param[in] input_x Width of input tensor
- * @param[in] input_y Height of input tensor
- * @param[in] input_ch Channels in input tensor
- * @param[in] kernel Pointer to kernel weights
- * @param[in] kernel_x Width of kernel
- * @param[in] kernel_y Height of kernel
- * @param[in] ch_mult Number of channel multiplier
- * @param[in] pad_x Padding sizes x
- * @param[in] pad_y Padding sizes y
- * @param[in] stride_x Convolution stride along the width
- * @param[in] stride_y Convolution stride along the height
- * @param[in] dilation_x Dilation along width. Not used and intended for future enhancement.
- * @param[in] dilation_y Dilation along height. Not used and intended for future enhancement.
- * @param[in] bias Pointer to optional bias values. If no bias is
- * available, NULL is expected
- * @param[in] input_offset Input tensor zero offset
- * @param[in] filter_offset Kernel tensor zero offset
- * @param[in] output_offset Output tensor zero offset
- * @param[in,out] output Pointer to output tensor
- * @param[in] output_x Width of output tensor
- * @param[in] output_y Height of output tensor
- * @param[in] output_activation_min Minimum value to clamp the output to. Range : {0, 255}
- * @param[in] output_activation_max Minimum value to clamp the output to. Range : {0, 255}
- * @param[in] out_shift Amount of right-shift for output
- * @param[in] out_mult Output multiplier for requantization
- * @return The function returns one of the following
- * RISCV_MATH_SIZE_MISMATCH - Not supported dimension of tensors
- * RISCV_MATH_SUCCESS - Successful operation
- * RISCV_MATH_ARGUMENT_ERROR - Implementation not available
- *
- * Input constraints
- * ch_mult is multiple of 2
- * kernel_x is multiple of 2
- *
- */
- riscv_status riscv_depthwise_conv_u8_basic_ver1(const uint8_t *input,
- const uint16_t input_x,
- const uint16_t input_y,
- const uint16_t input_ch,
- const uint8_t *kernel,
- const uint16_t kernel_x,
- const uint16_t kernel_y,
- const int16_t ch_mult,
- const int16_t pad_x,
- const int16_t pad_y,
- const int16_t stride_x,
- const int16_t stride_y,
- const int16_t dilation_x,
- const int16_t dilation_y,
- const int32_t *bias,
- const int32_t input_offset,
- const int32_t filter_offset,
- const int32_t output_offset,
- uint8_t *output,
- const uint16_t output_x,
- const uint16_t output_y,
- const int32_t output_activation_min,
- const int32_t output_activation_max,
- const int32_t out_shift,
- const int32_t out_mult);
-#ifdef __cplusplus
-}
-#endif
-
-#endif
diff --git a/arch/risc-v/nuclei/gcc/nmsis/NN/Include/riscv_nnsupportfunctions.h b/arch/risc-v/nuclei/gcc/nmsis/NN/Include/riscv_nnsupportfunctions.h
deleted file mode 100644
index 6061f08d..00000000
--- a/arch/risc-v/nuclei/gcc/nmsis/NN/Include/riscv_nnsupportfunctions.h
+++ /dev/null
@@ -1,366 +0,0 @@
-/*
- * Copyright (C) 2010-2018 Arm Limited or its affiliates. All rights reserved.
- * Copyright (c) 2019 Nuclei Limited. All rights reserved.
- *
- * SPDX-License-Identifier: Apache-2.0
- *
- * Licensed under the Apache License, Version 2.0 (the License); you may
- * not use this file except in compliance with the License.
- * You may obtain a copy of the License at
- *
- * www.apache.org/licenses/LICENSE-2.0
- *
- * Unless required by applicable law or agreed to in writing, software
- * distributed under the License is distributed on an AS IS BASIS, WITHOUT
- * WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
- * See the License for the specific language governing permissions and
- * limitations under the License.
- */
-
-/* ----------------------------------------------------------------------
- * Project: NMSIS NN Library
- * Title: riscv_nnsupportfunctions.h
- * Description: Public header file of support functions for NMSIS NN Library
- *
- * $Date: July 2019
- * $Revision: V.1.0.0
- *
- * Target Processor: RISC-V Cores
- * -------------------------------------------------------------------- */
-
-#ifndef _RISCV_NNSUPPORTFUNCTIONS_H_
-#define _RISCV_NNSUPPORTFUNCTIONS_H_
-
-#include "riscv_math.h"
-#include "riscv_common_tables.h"
-
-#ifdef __cplusplus
-extern "C"
-{
-#endif
-
-#define LEFT_SHIFT(_shift) (_shift > 0 ? _shift : 0)
-#define RIGHT_SHIFT(_shift) (_shift > 0 ? 0 : -_shift)
-#define Q31_MIN (0x80000000L)
-#define Q31_MAX (0x7FFFFFFFL)
-
-#define MAX(A,B) (A) > (B) ? (A) : (B)
-#define MIN(A,B) (A) < (B) ? (A) : (B)
-
-/**
- * @brief Union for SIMD access of q31/q15/q7 types
- */
-union riscv_nnword
-{
- q31_t word;
- /**< q31 type */
- q15_t half_words[2];
- /**< q15 type */
- q7_t bytes[4];
- /**< q7 type */
-};
-
-/**
- * @brief Struct for specifying activation function types
- *
- */
-typedef enum
-{
- RISCV_SIGMOID = 0,
- /**< Sigmoid activation function */
- RISCV_TANH = 1,
- /**< Tanh activation function */
-} riscv_nn_activation_type;
-
-/**
- * @defgroup nndata_convert Neural Network Data Conversion Functions
- *
- * Perform data type conversion in-between neural network operations
- *
- */
-
-/**
- * @brief Converts the elements of the q7 vector to q15 vector without left-shift
- * @param[in] *pSrc points to the q7 input vector
- * @param[out] *pDst points to the q15 output vector
- * @param[in] blockSize length of the input vector
- * @return none.
- *
- */
-void riscv_q7_to_q15_no_shift(const q7_t * pSrc, q15_t * pDst, uint32_t blockSize);
-
-void riscv_q7_to_q7_no_shift(const q7_t * pSrc, q7_t * pDst, uint32_t blockSize);
-
-/**
- * @brief Non-saturating addition of elements of a q7 vector
- * @param[in] *input Pointer to the q7 input vector
- * @param[out] *output Pointer to the q31 output variable.
- * @param[in] block_size length of the input vector
- * @return none.
- * \par Description:
- *
- * 2^24 samples can be added without saturating the result.
- *
- * The equation used for the conversion process is:
- *
- *
- * sum = input[0] + input[1] + .. + input[block_size -1]
- *
- *
- * */
-void riscv_nn_add_q7(const q7_t *input, q31_t *output, uint32_t block_size);
-
-/**
- * @brief Converts the elements of the q7 vector to reordered q15 vector without left-shift
- * @param[in] *pSrc points to the q7 input vector
- * @param[out] *pDst points to the q15 output vector
- * @param[in] blockSize length of the input vector
- * @return none.
- *
- */
-void riscv_q7_to_q15_reordered_no_shift(const q7_t * pSrc, q15_t * pDst, uint32_t blockSize);
-
-void riscv_q7_to_q7_reordered_no_shift(const q7_t * pSrc, q7_t * pDst, uint32_t blockSize);
-
-/**
- * @brief Converts the elements from a q7 vector to a q15 vector with an added offset
- * @param[in] *src points to the q7 input vector
- * @param[out] *dst points to the q15 output vector
- * @param[in] block_size length of the input vector
- * @param[in] offset q7 offset to be added to each input vector element.
- * @return none.
- *
- * \par Description:
- *
- * The equation used for the conversion process is:
- *
- *
- * dst[n] = (q15_t) src[n] + offset; 0 <= n < block_size.
- *
- *
- */
-void riscv_q7_to_q15_with_offset(const q7_t *src, q15_t *dst, uint32_t block_size, q7_t offset);
-
-#if defined (RISCV_MATH_DSP)
-
-/**
- * @brief read and expand one q7 word into two q15 words
- */
-
-__STATIC_FORCEINLINE void *read_and_pad(void *source, q31_t * out1, q31_t * out2)
-{
- q31_t inA = *__SIMD32(source)++;
- q31_t inAbuf1 = __SXTB16(__ROR(inA, 8));
- q31_t inAbuf2 = __SXTB16(inA);
-
- *out2 = __PKHTB(inAbuf1, inAbuf2, 16);
- *out1 = __PKHBT(inAbuf2, inAbuf1, 16);
-
- return source;
-}
-
-/**
- * @brief read and expand one q7 word into two q15 words with reordering
- */
-
-__STATIC_FORCEINLINE q7_t *read_and_pad_reordered(q7_t *source, q31_t * out1, q31_t * out2)
-{
- q31_t inA = read_q7x4_ia(&source);
- *out2 = __SXTB16(__ROR(inA, 8));
- *out1 = __SXTB16(inA);
-
- return source;
-}
-
-/**
- * @brief read and expand one q7 word into two q15 words with reordering and add an offset
- */
-__STATIC_FORCEINLINE q7_t *read_and_pad_reordered_with_offset(q7_t *source, q31_t * out1, q31_t * out2,q31_t offset)
-{
- q31_t inA = read_q7x4_ia(&source);
-
- *out2 = __SXTB16(__ROR(inA, 8));
- *out1 = __SXTB16(inA);
- *out1 = __QADD16(*out1,offset);
- *out2 = __QADD16(*out2,offset);
-
- return source;
-}
-
-
-#endif
-
-
-
-/**
- * @defgroup NNBasicMath Basic Math Functions for Neural Network Computation
- *
- * Basic Math Functions for Neural Network Computation
- *
- */
-
-/**
- * @brief q7 vector multiplication with variable output shifts
- * @param[in] *pSrcA pointer to the first input vector
- * @param[in] *pSrcB pointer to the second input vector
- * @param[out] *pDst pointer to the output vector
- * @param[in] out_shift amount of right-shift for output
- * @param[in] blockSize number of samples in each vector
- * @return none.
- *
- * Scaling and Overflow Behavior:
- * \par
- * The function uses saturating arithmetic.
- * Results outside of the allowable q15 range [0x8000 0x7FFF] will be saturated.
- */
-
-void riscv_nn_mult_q15(
- q15_t * pSrcA,
- q15_t * pSrcB,
- q15_t * pDst,
- const uint16_t out_shift,
- uint32_t blockSize);
-
-/**
- * @brief q7 vector multiplication with variable output shifts
- * @param[in] *pSrcA pointer to the first input vector
- * @param[in] *pSrcB pointer to the second input vector
- * @param[out] *pDst pointer to the output vector
- * @param[in] out_shift amount of right-shift for output
- * @param[in] blockSize number of samples in each vector
- * @return none.
- *
- * Scaling and Overflow Behavior:
- * \par
- * The function uses saturating arithmetic.
- * Results outside of the allowable q7 range [0x80 0x7F] will be saturated.
- */
-
-void riscv_nn_mult_q7(
- q7_t * pSrcA,
- q7_t * pSrcB,
- q7_t * pDst,
- const uint16_t out_shift,
- uint32_t blockSize);
-
-/**
- * @brief macro for adding rounding offset
- */
-#ifndef RISCV_NN_TRUNCATE
- #define NN_ROUND(out_shift) ( (0x1 << out_shift) >> 1 )
-#else
- #define NN_ROUND(out_shift) 0
-#endif
-
-/**
- * @brief Saturating doubling high multiply. Result matches
- * NEON instruction VQRDMULH.
- * @param[in] m1 Multiplicand
- * @param[in] m2 Multiplier
- * @return Result of multiplication.
- *
- */
-__STATIC_FORCEINLINE q31_t riscv_nn_sat_doubling_high_mult(const q31_t m1, const q31_t m2)
-{
- q31_t result = 0;
- // Rounding offset to add for a right shift of 31
- q63_t mult = 1 << 30;
-
- if ((m1 < 0) ^ (m2 < 0))
- {
- mult = 1 - mult;
- }
- // Gets resolved as a SMLAL instruction
- mult = mult + (q63_t)m1 * m2;
-
- // Utilize all of the upper 32 bits. This is the doubling step
- // as well.
- result = mult / (1UL << 31);
-
- if ((m1 == m2) && (m1 == Q31_MIN))
- {
- result = Q31_MAX;
- }
- return result;
-}
-
-/**
- * @brief Rounding divide by power of two.
- * @param[in] dividend - Dividend
- * @param[in] exponent - Divisor = power(2, exponent)
- * Range: [0, 31]
- * @return Rounded result of division. Midpoint is rounded away from zero.
- *
- */
-__STATIC_FORCEINLINE q31_t riscv_nn_divide_by_power_of_two(const q31_t dividend, const q31_t exponent)
-{
- q31_t result = 0;
- const q31_t remainder_mask = (1l << exponent) - 1;
- int32_t remainder = remainder_mask & dividend;
-
- // Basic division
- result = dividend >> exponent;
-
- // Adjust 'result' for rounding (mid point away from zero)
- q31_t threshold = remainder_mask >> 1;
- if (result < 0)
- {
- threshold++;
- }
- if (remainder > threshold)
- {
- result++;
- }
-
- return result;
-}
-
-/**
- * @brief Requantize a given value.
- * @param[in] val Value to be requantized
- * @param[in] multiplier multiplier
- * @param[in] shift left or right shift for 'val * multiplier'
- *
- * @return Returns (val * multiplier)/(2 ^ shift)
- *
- */
-__STATIC_FORCEINLINE q31_t riscv_nn_requantize(const q31_t val, const q31_t multiplier, const q31_t shift)
-{
- return riscv_nn_divide_by_power_of_two(riscv_nn_sat_doubling_high_mult(val * (1 << LEFT_SHIFT(shift)), multiplier),
- RIGHT_SHIFT(shift));
-}
-
-/**
- @brief Read 2 q15 elements and post increment pointer.
- @param[in] in_q15 Pointer to pointer that holds address of input.
- @return q31 value
- */
-__STATIC_FORCEINLINE q31_t riscv_nn_read_q15x2_ia(const q15_t **in_q15)
-{
- q31_t val;
-
- memcpy(&val, *in_q15, 4);
- *in_q15 += 2;
-
- return (val);
-}
-
-/**
- @brief Read 4 q7 from q7 pointer and post increment pointer.
- @param[in] in_q7 Pointer to pointer that holds address of input.
- @return q31 value
- */
-__STATIC_FORCEINLINE q31_t riscv_nn_read_q7x4_ia(const q7_t **in_q7)
-{
- q31_t val;
- memcpy(&val, *in_q7, 4);
- *in_q7 += 4;
-
- return (val);
-}
-
-#ifdef __cplusplus
-}
-#endif
-
-#endif