CMSIS-DSP: Improvements to the float16 support for building.
Christophe Favergeon
6 years ago
parent
1464ad53f7
commit
c33adc9485
@ -0,0 +1,90 @@
|
||||
/* ----------------------------------------------------------------------
|
||||
* Project: CMSIS DSP Library
|
||||
* Title: arm_common_tables_f16.h
|
||||
* Description: Extern declaration for common tables
|
||||
*
|
||||
* $Date: 27. January 2017
|
||||
* $Revision: V.1.5.1
|
||||
*
|
||||
* Target Processor: Cortex-M cores
|
||||
* -------------------------------------------------------------------- */
|
||||
/*
|
||||
* Copyright (C) 2010-2017 ARM Limited or its affiliates. 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 _ARM_COMMON_TABLES_F16_H
|
||||
#define _ARM_COMMON_TABLES_F16_H
|
||||
|
||||
#include "arm_math_f16.h"
|
||||
#include "arm_common_tables.h"
|
||||
|
||||
#ifdef __cplusplus
|
||||
extern "C"
|
||||
{
|
||||
#endif
|
||||
|
||||
#if !defined(ARM_DSP_CONFIG_TABLES) || defined(ARM_FFT_ALLOW_TABLES)
|
||||
|
||||
/* F16 */
|
||||
#if !defined(__CC_ARM) && defined(ARM_FLOAT16_SUPPORTED)
|
||||
#if !defined(ARM_DSP_CONFIG_TABLES) || defined(ARM_ALL_FFT_TABLES) || defined(ARM_TABLE_TWIDDLECOEF_F16_16)
|
||||
extern const float16_t twiddleCoefF16_16[32];
|
||||
#endif /* !defined(ARM_DSP_CONFIG_TABLES) || defined(ARM_ALL_FFT_TABLES) */
|
||||
|
||||
#if !defined(ARM_DSP_CONFIG_TABLES) || defined(ARM_ALL_FFT_TABLES) || defined(ARM_TABLE_TWIDDLECOEF_F16_32)
|
||||
extern const float16_t twiddleCoefF16_32[64];
|
||||
#endif /* !defined(ARM_DSP_CONFIG_TABLES) || defined(ARM_ALL_FFT_TABLES) */
|
||||
|
||||
#if !defined(ARM_DSP_CONFIG_TABLES) || defined(ARM_ALL_FFT_TABLES) || defined(ARM_TABLE_TWIDDLECOEF_F16_64)
|
||||
extern const float16_t twiddleCoefF16_64[128];
|
||||
#endif /* !defined(ARM_DSP_CONFIG_TABLES) || defined(ARM_ALL_FFT_TABLES) */
|
||||
|
||||
#if !defined(ARM_DSP_CONFIG_TABLES) || defined(ARM_ALL_FFT_TABLES) || defined(ARM_TABLE_TWIDDLECOEF_F16_128)
|
||||
extern const float16_t twiddleCoefF16_128[256];
|
||||
#endif /* !defined(ARM_DSP_CONFIG_TABLES) || defined(ARM_ALL_FFT_TABLES) */
|
||||
|
||||
#if !defined(ARM_DSP_CONFIG_TABLES) || defined(ARM_ALL_FFT_TABLES) || defined(ARM_TABLE_TWIDDLECOEF_F16_256)
|
||||
extern const float16_t twiddleCoefF16_256[512];
|
||||
#endif /* !defined(ARM_DSP_CONFIG_TABLES) || defined(ARM_ALL_FFT_TABLES) */
|
||||
|
||||
#if !defined(ARM_DSP_CONFIG_TABLES) || defined(ARM_ALL_FFT_TABLES) || defined(ARM_TABLE_TWIDDLECOEF_F16_512)
|
||||
extern const float16_t twiddleCoefF16_512[1024];
|
||||
#endif /* !defined(ARM_DSP_CONFIG_TABLES) || defined(ARM_ALL_FFT_TABLES) */
|
||||
|
||||
#if !defined(ARM_DSP_CONFIG_TABLES) || defined(ARM_ALL_FFT_TABLES) || defined(ARM_TABLE_TWIDDLECOEF_F16_1024)
|
||||
extern const float16_t twiddleCoefF16_1024[2048];
|
||||
#endif /* !defined(ARM_DSP_CONFIG_TABLES) || defined(ARM_ALL_FFT_TABLES) */
|
||||
|
||||
#if !defined(ARM_DSP_CONFIG_TABLES) || defined(ARM_ALL_FFT_TABLES) || defined(ARM_TABLE_TWIDDLECOEF_F16_2048)
|
||||
extern const float16_t twiddleCoefF16_2048[4096];
|
||||
#endif /* !defined(ARM_DSP_CONFIG_TABLES) || defined(ARM_ALL_FFT_TABLES) */
|
||||
|
||||
#if !defined(ARM_DSP_CONFIG_TABLES) || defined(ARM_ALL_FFT_TABLES) || defined(ARM_TABLE_TWIDDLECOEF_F16_4096)
|
||||
extern const float16_t twiddleCoefF16_4096[8192];
|
||||
#endif /* !defined(ARM_DSP_CONFIG_TABLES) || defined(ARM_ALL_FFT_TABLES) */
|
||||
#endif /* ARMAC5 */
|
||||
|
||||
#endif /* !defined(ARM_DSP_CONFIG_TABLES) || defined(ARM_FFT_ALLOW_TABLES) */
|
||||
|
||||
|
||||
#ifdef __cplusplus
|
||||
}
|
||||
#endif
|
||||
|
||||
#endif /* _ARM_COMMON_TABLES_F16_H */
|
||||
|
||||
|
||||
@ -0,0 +1,57 @@
|
||||
/* ----------------------------------------------------------------------
|
||||
* Project: CMSIS DSP Library
|
||||
* Title: arm_const_structs_f16.h
|
||||
* Description: Constant structs that are initialized for user convenience.
|
||||
* For example, some can be given as arguments to the arm_cfft_f16() function.
|
||||
*
|
||||
* $Date: 20. April 2020
|
||||
* $Revision: V.1.5.1
|
||||
*
|
||||
* Target Processor: Cortex-M cores
|
||||
* -------------------------------------------------------------------- */
|
||||
/*
|
||||
* Copyright (C) 2010-2020 ARM Limited or its affiliates. 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 _ARM_CONST_STRUCTS_F16_H
|
||||
#define _ARM_CONST_STRUCTS_F16_H
|
||||
|
||||
#include "arm_math_f16.h"
|
||||
#include "arm_common_tables_f16.h"
|
||||
|
||||
#ifdef __cplusplus
|
||||
extern "C"
|
||||
{
|
||||
#endif
|
||||
|
||||
#if !defined(__CC_ARM) && defined(ARM_FLOAT16_SUPPORTED)
|
||||
extern const arm_cfft_instance_f16 arm_cfft_sR_f16_len16;
|
||||
extern const arm_cfft_instance_f16 arm_cfft_sR_f16_len32;
|
||||
extern const arm_cfft_instance_f16 arm_cfft_sR_f16_len64;
|
||||
extern const arm_cfft_instance_f16 arm_cfft_sR_f16_len128;
|
||||
extern const arm_cfft_instance_f16 arm_cfft_sR_f16_len256;
|
||||
extern const arm_cfft_instance_f16 arm_cfft_sR_f16_len512;
|
||||
extern const arm_cfft_instance_f16 arm_cfft_sR_f16_len1024;
|
||||
extern const arm_cfft_instance_f16 arm_cfft_sR_f16_len2048;
|
||||
extern const arm_cfft_instance_f16 arm_cfft_sR_f16_len4096;
|
||||
#endif
|
||||
|
||||
#ifdef __cplusplus
|
||||
}
|
||||
#endif
|
||||
|
||||
#endif
|
||||
@ -0,0 +1,310 @@
|
||||
/******************************************************************************
|
||||
* @file arm_math_f16.h
|
||||
* @brief Public header file for f16 function of the CMSIS DSP Library
|
||||
* @version V1.8.1
|
||||
* @date 20. April 2020
|
||||
******************************************************************************/
|
||||
/*
|
||||
* Copyright (c) 2010-2020 Arm Limited or its affiliates. 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 _ARM_MATH_F16_H
|
||||
#define _ARM_MATH_F16_H
|
||||
|
||||
#include "arm_math.h"
|
||||
|
||||
#ifdef __cplusplus
|
||||
extern "C"
|
||||
{
|
||||
#endif
|
||||
|
||||
#if !defined( __CC_ARM )
|
||||
|
||||
/**
|
||||
* @brief 16-bit floating-point type definition.
|
||||
* This is already defined in arm_mve.h
|
||||
*
|
||||
* This is not fully supported on ARM AC5.
|
||||
*/
|
||||
|
||||
/*
|
||||
|
||||
Check if the type __fp16 is available.
|
||||
If it is not available, f16 version of the kernels
|
||||
won't be built.
|
||||
|
||||
*/
|
||||
#if !(__ARM_FEATURE_MVE & 2) && !(__ARM_NEON)
|
||||
#if defined(__ARM_FP16_FORMAT_IEEE) || defined(__ARM_FP16_FORMAT_ALTERNATIVE)
|
||||
typedef __fp16 float16_t;
|
||||
#define ARM_FLOAT16_SUPPORTED
|
||||
#endif
|
||||
#else
|
||||
#define ARM_FLOAT16_SUPPORTED
|
||||
#endif
|
||||
|
||||
#if defined(ARM_MATH_NEON) || defined(ARM_MATH_MVEF) /* floating point vector*/
|
||||
|
||||
#if defined(ARM_MATH_MVE_FLOAT16) || defined(ARM_MATH_NEON_FLOAT16)
|
||||
/**
|
||||
* @brief 16-bit floating-point 128-bit vector data type
|
||||
*/
|
||||
typedef __ALIGNED(2) float16x8_t f16x8_t;
|
||||
|
||||
/**
|
||||
* @brief 16-bit floating-point 128-bit vector pair data type
|
||||
*/
|
||||
typedef float16x8x2_t f16x8x2_t;
|
||||
|
||||
/**
|
||||
* @brief 16-bit floating-point 128-bit vector quadruplet data type
|
||||
*/
|
||||
typedef float16x8x4_t f16x8x4_t;
|
||||
|
||||
/**
|
||||
* @brief 16-bit ubiquitous 128-bit vector data type
|
||||
*/
|
||||
typedef union _any16x8_t
|
||||
{
|
||||
float16x8_t f;
|
||||
int16x8_t i;
|
||||
} any16x8_t;
|
||||
#endif
|
||||
|
||||
#endif
|
||||
|
||||
#if defined(ARM_MATH_NEON)
|
||||
|
||||
|
||||
#if defined(ARM_MATH_NEON_FLOAT16)
|
||||
/**
|
||||
* @brief 16-bit float 64-bit vector data type.
|
||||
*/
|
||||
typedef __ALIGNED(2) float16x4_t f16x4_t;
|
||||
|
||||
/**
|
||||
* @brief 16-bit floating-point 128-bit vector triplet data type
|
||||
*/
|
||||
typedef float16x8x3_t f16x8x3_t;
|
||||
|
||||
/**
|
||||
* @brief 16-bit floating-point 64-bit vector pair data type
|
||||
*/
|
||||
typedef float16x4x2_t f16x4x2_t;
|
||||
|
||||
/**
|
||||
* @brief 16-bit floating-point 64-bit vector triplet data type
|
||||
*/
|
||||
typedef float16x4x3_t f16x4x3_t;
|
||||
|
||||
/**
|
||||
* @brief 16-bit floating-point 64-bit vector quadruplet data type
|
||||
*/
|
||||
typedef float16x4x4_t f16x4x4_t;
|
||||
|
||||
/**
|
||||
* @brief 16-bit ubiquitous 64-bit vector data type
|
||||
*/
|
||||
typedef union _any16x4_t
|
||||
{
|
||||
float16x4_t f;
|
||||
int16x4_t i;
|
||||
} any16x4_t;
|
||||
#endif
|
||||
|
||||
#endif
|
||||
|
||||
|
||||
|
||||
#if defined(ARM_FLOAT16_SUPPORTED)
|
||||
#define F16_MAX ((float16_t)FLT_MAX)
|
||||
#define F16_MIN (-(float16_t)FLT_MAX)
|
||||
|
||||
#define F16_ABSMAX ((float16_t)FLT_MAX)
|
||||
#define F16_ABSMIN ((float16_t)0.0)
|
||||
|
||||
/**
|
||||
* @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 arm_add_f16(
|
||||
const float16_t * pSrcA,
|
||||
const float16_t * pSrcB,
|
||||
float16_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 arm_sub_f16(
|
||||
const float16_t * pSrcA,
|
||||
const float16_t * pSrcB,
|
||||
float16_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 arm_scale_f16(
|
||||
const float16_t * pSrc,
|
||||
float16_t scale,
|
||||
float16_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 arm_abs_f16(
|
||||
const float16_t * pSrc,
|
||||
float16_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 arm_offset_f16(
|
||||
const float16_t * pSrc,
|
||||
float16_t offset,
|
||||
float16_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 arm_dot_prod_f16(
|
||||
const float16_t * pSrcA,
|
||||
const float16_t * pSrcB,
|
||||
uint32_t blockSize,
|
||||
float16_t * result);
|
||||
|
||||
/**
|
||||
* @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 arm_mult_f16(
|
||||
const float16_t * pSrcA,
|
||||
const float16_t * pSrcB,
|
||||
float16_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 arm_negate_f16(
|
||||
const float16_t * pSrc,
|
||||
float16_t * pDst,
|
||||
uint32_t blockSize);
|
||||
|
||||
/**
|
||||
* @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 float16_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. */
|
||||
float16_t onebyfftLen; /**< value of 1/fftLen. */
|
||||
} arm_cfft_radix2_instance_f16;
|
||||
|
||||
/**
|
||||
* @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 float16_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. */
|
||||
float16_t onebyfftLen; /**< value of 1/fftLen. */
|
||||
} arm_cfft_radix4_instance_f16;
|
||||
|
||||
/**
|
||||
* @brief Instance structure for the floating-point CFFT/CIFFT function.
|
||||
*/
|
||||
typedef struct
|
||||
{
|
||||
uint16_t fftLen; /**< length of the FFT. */
|
||||
const float16_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. */
|
||||
#if defined(ARM_MATH_MVEF) && !defined(ARM_MATH_AUTOVECTORIZE)
|
||||
const uint32_t *rearranged_twiddle_tab_stride1_arr; /**< Per stage reordered twiddle pointer (offset 1) */ \
|
||||
const uint32_t *rearranged_twiddle_tab_stride2_arr; /**< Per stage reordered twiddle pointer (offset 2) */ \
|
||||
const uint32_t *rearranged_twiddle_tab_stride3_arr; /**< Per stage reordered twiddle pointer (offset 3) */ \
|
||||
const float16_t *rearranged_twiddle_stride1; /**< reordered twiddle offset 1 storage */ \
|
||||
const float16_t *rearranged_twiddle_stride2; /**< reordered twiddle offset 2 storage */ \
|
||||
const float16_t *rearranged_twiddle_stride3;
|
||||
#endif
|
||||
} arm_cfft_instance_f16;
|
||||
|
||||
|
||||
arm_status arm_cfft_init_f16(
|
||||
arm_cfft_instance_f16 * S,
|
||||
uint16_t fftLen);
|
||||
|
||||
void arm_cfft_f16(
|
||||
const arm_cfft_instance_f16 * S,
|
||||
float16_t * p1,
|
||||
uint8_t ifftFlag,
|
||||
uint8_t bitReverseFlag);
|
||||
|
||||
#endif /* ARM_FLOAT16_SUPPORTED*/
|
||||
#endif /* !defined( __CC_ARM ) */
|
||||
|
||||
#ifdef __cplusplus
|
||||
}
|
||||
#endif
|
||||
|
||||
#endif /* _ARM_MATH_F16_H */
|
||||
|
||||
|
||||
@ -0,0 +1,108 @@
|
||||
/* ----------------------------------------------------------------------
|
||||
* Project: CMSIS DSP Library
|
||||
* Title: arm_mve_tables_f16.h
|
||||
* Description: common tables like fft twiddle factors, Bitreverse, reciprocal etc
|
||||
* used for MVE implementation only
|
||||
*
|
||||
* $Date: 14. April 2020
|
||||
*
|
||||
* Target Processor: Cortex-M cores
|
||||
* -------------------------------------------------------------------- */
|
||||
/*
|
||||
* Copyright (C) 2010-2020 ARM Limited or its affiliates. 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 _ARM_MVE_TABLES_F16_H
|
||||
#define _ARM_MVE_TABLES_F16_H
|
||||
|
||||
#include "arm_math_f16.h"
|
||||
|
||||
#ifdef __cplusplus
|
||||
extern "C"
|
||||
{
|
||||
#endif
|
||||
|
||||
|
||||
|
||||
|
||||
#if defined(ARM_MATH_MVEF) && !defined(ARM_MATH_AUTOVECTORIZE)
|
||||
|
||||
#if !defined(ARM_DSP_CONFIG_TABLES) || defined(ARM_FFT_ALLOW_TABLES)
|
||||
|
||||
#if !defined(ARM_DSP_CONFIG_TABLES) || defined(ARM_ALL_FFT_TABLES) || defined(ARM_TABLE_TWIDDLECOEF_F16_16) || defined(ARM_TABLE_TWIDDLECOEF_F16_32)
|
||||
|
||||
extern uint32_t rearranged_twiddle_tab_stride1_arr_16_f16[2];
|
||||
extern uint32_t rearranged_twiddle_tab_stride2_arr_16_f16[2];
|
||||
extern uint32_t rearranged_twiddle_tab_stride3_arr_16_f16[2];
|
||||
extern float16_t rearranged_twiddle_stride1_16_f16[8];
|
||||
extern float16_t rearranged_twiddle_stride2_16_f16[8];
|
||||
extern float16_t rearranged_twiddle_stride3_16_f16[8];
|
||||
#endif
|
||||
|
||||
#if !defined(ARM_DSP_CONFIG_TABLES) || defined(ARM_ALL_FFT_TABLES) || defined(ARM_TABLE_TWIDDLECOEF_F16_64) || defined(ARM_TABLE_TWIDDLECOEF_F16_128)
|
||||
|
||||
extern uint32_t rearranged_twiddle_tab_stride1_arr_64_f16[3];
|
||||
extern uint32_t rearranged_twiddle_tab_stride2_arr_64_f16[3];
|
||||
extern uint32_t rearranged_twiddle_tab_stride3_arr_64_f16[3];
|
||||
extern float16_t rearranged_twiddle_stride1_64_f16[40];
|
||||
extern float16_t rearranged_twiddle_stride2_64_f16[40];
|
||||
extern float16_t rearranged_twiddle_stride3_64_f16[40];
|
||||
#endif
|
||||
|
||||
#if !defined(ARM_DSP_CONFIG_TABLES) || defined(ARM_ALL_FFT_TABLES) || defined(ARM_TABLE_TWIDDLECOEF_F16_256) || defined(ARM_TABLE_TWIDDLECOEF_F16_512)
|
||||
|
||||
extern uint32_t rearranged_twiddle_tab_stride1_arr_256_f16[4];
|
||||
extern uint32_t rearranged_twiddle_tab_stride2_arr_256_f16[4];
|
||||
extern uint32_t rearranged_twiddle_tab_stride3_arr_256_f16[4];
|
||||
extern float16_t rearranged_twiddle_stride1_256_f16[168];
|
||||
extern float16_t rearranged_twiddle_stride2_256_f16[168];
|
||||
extern float16_t rearranged_twiddle_stride3_256_f16[168];
|
||||
#endif
|
||||
|
||||
#if !defined(ARM_DSP_CONFIG_TABLES) || defined(ARM_ALL_FFT_TABLES) || defined(ARM_TABLE_TWIDDLECOEF_F16_1024) || defined(ARM_TABLE_TWIDDLECOEF_F16_2048)
|
||||
|
||||
extern uint32_t rearranged_twiddle_tab_stride1_arr_1024_f16[5];
|
||||
extern uint32_t rearranged_twiddle_tab_stride2_arr_1024_f16[5];
|
||||
extern uint32_t rearranged_twiddle_tab_stride3_arr_1024_f16[5];
|
||||
extern float16_t rearranged_twiddle_stride1_1024_f16[680];
|
||||
extern float16_t rearranged_twiddle_stride2_1024_f16[680];
|
||||
extern float16_t rearranged_twiddle_stride3_1024_f16[680];
|
||||
#endif
|
||||
|
||||
#if !defined(ARM_DSP_CONFIG_TABLES) || defined(ARM_ALL_FFT_TABLES) || defined(ARM_TABLE_TWIDDLECOEF_F16_4096) || defined(ARM_TABLE_TWIDDLECOEF_F16_8192)
|
||||
|
||||
extern uint32_t rearranged_twiddle_tab_stride1_arr_4096_f16[6];
|
||||
extern uint32_t rearranged_twiddle_tab_stride2_arr_4096_f16[6];
|
||||
extern uint32_t rearranged_twiddle_tab_stride3_arr_4096_f16[6];
|
||||
extern float16_t rearranged_twiddle_stride1_4096_f16[2728];
|
||||
extern float16_t rearranged_twiddle_stride2_4096_f16[2728];
|
||||
extern float16_t rearranged_twiddle_stride3_4096_f16[2728];
|
||||
#endif
|
||||
|
||||
|
||||
#endif /* !defined(ARM_DSP_CONFIG_TABLES) || defined(ARM_FFT_ALLOW_TABLES) */
|
||||
|
||||
#endif /* defined(ARM_MATH_MVEF) && !defined(ARM_MATH_AUTOVECTORIZE) */
|
||||
|
||||
|
||||
|
||||
#ifdef __cplusplus
|
||||
}
|
||||
#endif
|
||||
|
||||
#endif /*_ARM_MVE_TABLES_F16_H*/
|
||||
|
||||
@ -0,0 +1,36 @@
|
||||
/* ----------------------------------------------------------------------
|
||||
* Project: CMSIS DSP Library
|
||||
* Title: BasicMathFunctionsF16.c
|
||||
* Description: Combination of all basic math function f16 source files.
|
||||
*
|
||||
* $Date: 20. April 2020
|
||||
* $Revision: V1.1.0
|
||||
*
|
||||
* Target Processor: Cortex-M cores
|
||||
* -------------------------------------------------------------------- */
|
||||
/*
|
||||
* Copyright (C) 2019-2020 ARM Limited or its affiliates. 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.
|
||||
*/
|
||||
|
||||
#include "arm_abs_f16.c"
|
||||
#include "arm_add_f16.c"
|
||||
#include "arm_dot_prod_f16.c"
|
||||
#include "arm_mult_f16.c"
|
||||
#include "arm_negate_f16.c"
|
||||
#include "arm_offset_f16.c"
|
||||
#include "arm_scale_f16.c"
|
||||
#include "arm_sub_f16.c"
|
||||
@ -0,0 +1,31 @@
|
||||
/* ----------------------------------------------------------------------
|
||||
* Project: CMSIS DSP Library
|
||||
* Title: CommonTables.c
|
||||
* Description: Combination of all common table source files.
|
||||
*
|
||||
* $Date: 08. January 2020
|
||||
* $Revision: V1.1.0
|
||||
*
|
||||
* Target Processor: Cortex-M cores
|
||||
* -------------------------------------------------------------------- */
|
||||
/*
|
||||
* Copyright (C) 2019-2020 ARM Limited or its affiliates. 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.
|
||||
*/
|
||||
|
||||
#include "arm_common_tables_f16.c"
|
||||
#include "arm_const_structs_f16.c"
|
||||
#include "arm_mve_tables_f16.c"
|
||||
File diff suppressed because it is too large
Load Diff
File diff suppressed because it is too large
Load Diff
@ -0,0 +1,101 @@
|
||||
/* ----------------------------------------------------------------------
|
||||
* Project: CMSIS DSP Library
|
||||
* Title: arm_const_structs_f16.c
|
||||
* Description: Constant structs that are initialized for user convenience.
|
||||
* For example, some can be given as arguments to the arm_cfft_f32() or arm_rfft_f32() functions.
|
||||
*
|
||||
* $Date: 27. January 2017
|
||||
* $Revision: V.1.5.1
|
||||
*
|
||||
* Target Processor: Cortex-M cores
|
||||
* -------------------------------------------------------------------- */
|
||||
/*
|
||||
* Copyright (C) 2010-2017 ARM Limited or its affiliates. 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.
|
||||
*/
|
||||
|
||||
#include "arm_math_f16.h"
|
||||
|
||||
#if defined(ARM_FLOAT16_SUPPORTED)
|
||||
|
||||
#include "arm_const_structs_f16.h"
|
||||
|
||||
|
||||
/*
|
||||
ALLOW TABLE is true when config table is enabled and the Tramsform folder is included
|
||||
for compilation.
|
||||
*/
|
||||
#if !defined(ARM_DSP_CONFIG_TABLES) || defined(ARM_FFT_ALLOW_TABLES)
|
||||
|
||||
|
||||
/* Floating-point structs */
|
||||
#if !defined(ARM_MATH_MVEF) || defined(ARM_MATH_AUTOVECTORIZE)
|
||||
|
||||
|
||||
/*
|
||||
|
||||
Those structures cannot be used to initialize the MVE version of the FFT F32 instances.
|
||||
So they are not compiled when MVE is defined.
|
||||
|
||||
For the MVE version, the new arm_cfft_init_f32 must be used.
|
||||
|
||||
|
||||
*/
|
||||
|
||||
#if !defined(__CC_ARM)
|
||||
const arm_cfft_instance_f16 arm_cfft_sR_f16_len16 = {
|
||||
16, twiddleCoefF16_16, armBitRevIndexTable_fixed_16, ARMBITREVINDEXTABLE_FIXED_16_TABLE_LENGTH
|
||||
};
|
||||
|
||||
const arm_cfft_instance_f16 arm_cfft_sR_f16_len32 = {
|
||||
32, twiddleCoefF16_32, armBitRevIndexTable_fixed_32, ARMBITREVINDEXTABLE_FIXED_32_TABLE_LENGTH
|
||||
};
|
||||
|
||||
const arm_cfft_instance_f16 arm_cfft_sR_f16_len64 = {
|
||||
64, twiddleCoefF16_64, armBitRevIndexTable_fixed_64, ARMBITREVINDEXTABLE_FIXED_64_TABLE_LENGTH
|
||||
};
|
||||
|
||||
const arm_cfft_instance_f16 arm_cfft_sR_f16_len128 = {
|
||||
128, twiddleCoefF16_128, armBitRevIndexTable_fixed_128, ARMBITREVINDEXTABLE_FIXED_128_TABLE_LENGTH
|
||||
};
|
||||
|
||||
const arm_cfft_instance_f16 arm_cfft_sR_f16_len256 = {
|
||||
256, twiddleCoefF16_256, armBitRevIndexTable_fixed_256, ARMBITREVINDEXTABLE_FIXED_256_TABLE_LENGTH
|
||||
};
|
||||
|
||||
const arm_cfft_instance_f16 arm_cfft_sR_f16_len512 = {
|
||||
512, twiddleCoefF16_512, armBitRevIndexTable_fixed_512, ARMBITREVINDEXTABLE_FIXED_512_TABLE_LENGTH
|
||||
};
|
||||
|
||||
const arm_cfft_instance_f16 arm_cfft_sR_f16_len1024 = {
|
||||
1024, twiddleCoefF16_1024, armBitRevIndexTable_fixed_1024, ARMBITREVINDEXTABLE_FIXED_1024_TABLE_LENGTH
|
||||
};
|
||||
|
||||
const arm_cfft_instance_f16 arm_cfft_sR_f16_len2048 = {
|
||||
2048, twiddleCoefF16_2048, armBitRevIndexTable_fixed_2048, ARMBITREVINDEXTABLE_FIXED_2048_TABLE_LENGTH
|
||||
};
|
||||
|
||||
const arm_cfft_instance_f16 arm_cfft_sR_f16_len4096 = {
|
||||
4096, twiddleCoefF16_4096, armBitRevIndexTable_fixed_4096, ARMBITREVINDEXTABLE_FIXED_4096_TABLE_LENGTH
|
||||
};
|
||||
#endif
|
||||
|
||||
#endif /* !defined(ARM_MATH_MVEF) || defined(ARM_MATH_AUTOVECTORIZE) */
|
||||
|
||||
|
||||
#endif
|
||||
|
||||
#endif
|
||||
File diff suppressed because it is too large
Load Diff
File diff suppressed because it is too large
Load Diff
@ -0,0 +1,33 @@
|
||||
/* ----------------------------------------------------------------------
|
||||
* Project: CMSIS DSP Library
|
||||
* Title: TransformFunctionsF16.c
|
||||
* Description: Combination of all transform function f16 source files.
|
||||
*
|
||||
* $Date: 20. April 2020
|
||||
* $Revision: V1.0.0
|
||||
*
|
||||
* Target Processor: Cortex-M cores
|
||||
* -------------------------------------------------------------------- */
|
||||
/*
|
||||
* Copyright (C) 2019 ARM Limited or its affiliates. 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.
|
||||
*/
|
||||
|
||||
#include "arm_cfft_f16.c"
|
||||
#include "arm_cfft_init_f16.c"
|
||||
#include "arm_cfft_radix2_f16.c"
|
||||
#include "arm_cfft_radix4_f16.c"
|
||||
|
||||
Loading…
Reference in New Issue