From 9c59a184dba820975e5da6fcd5d248aee87f7e2f Mon Sep 17 00:00:00 2001 From: Clyne Sullivan Date: Sun, 2 Feb 2025 11:26:53 -0500 Subject: add l476 implementation --- .../DSP/Source/FastMathFunctions/CMakeLists.txt | 131 ++++------ .../Source/FastMathFunctions/FastMathFunctions.c | 112 +++------ .../DSP/Source/FastMathFunctions/arm_cos_f32.c | 243 ++++++++++--------- .../DSP/Source/FastMathFunctions/arm_cos_q15.c | 168 ++++++------- .../DSP/Source/FastMathFunctions/arm_cos_q31.c | 168 ++++++------- .../DSP/Source/FastMathFunctions/arm_sin_f32.c | 244 +++++++++---------- .../DSP/Source/FastMathFunctions/arm_sin_q15.c | 160 ++++++------ .../DSP/Source/FastMathFunctions/arm_sin_q31.c | 159 ++++++------ .../DSP/Source/FastMathFunctions/arm_sqrt_q15.c | 267 ++++++++++---------- .../DSP/Source/FastMathFunctions/arm_sqrt_q31.c | 270 +++++++++++---------- 10 files changed, 942 insertions(+), 980 deletions(-) (limited to 'Drivers/CMSIS/DSP/Source/FastMathFunctions') diff --git a/Drivers/CMSIS/DSP/Source/FastMathFunctions/CMakeLists.txt b/Drivers/CMSIS/DSP/Source/FastMathFunctions/CMakeLists.txt index 95a9eb6..c42a05b 100644 --- a/Drivers/CMSIS/DSP/Source/FastMathFunctions/CMakeLists.txt +++ b/Drivers/CMSIS/DSP/Source/FastMathFunctions/CMakeLists.txt @@ -1,80 +1,51 @@ -cmake_minimum_required (VERSION 3.14) - -project(CMSISDSPFastMath) - -include(configLib) -include(configDsp) - - -add_library(CMSISDSPFastMath STATIC) -configLib(CMSISDSPFastMath ${ROOT}) -configDsp(CMSISDSPFastMath ${ROOT}) - -include(interpol) -interpol(CMSISDSPFastMath) - -if (CONFIGTABLE AND ALLFAST) - target_compile_definitions(CMSISDSPFastMath PUBLIC ARM_ALL_FAST_TABLES) -endif() - -if (NOT CONFIGTABLE OR ALLFAST OR ARM_COS_F32) -target_sources(CMSISDSPFastMath PRIVATE arm_cos_f32.c) -endif() - -if (NOT CONFIGTABLE OR ALLFAST OR ARM_COS_Q15) -target_sources(CMSISDSPFastMath PRIVATE arm_cos_q15.c) -endif() - -if (NOT CONFIGTABLE OR ALLFAST OR ARM_COS_Q31) -target_sources(CMSISDSPFastMath PRIVATE arm_cos_q31.c) -endif() - -if (NOT CONFIGTABLE OR ALLFAST OR ARM_SIN_F32) -target_sources(CMSISDSPFastMath PRIVATE arm_sin_f32.c) -endif() - -if (NOT CONFIGTABLE OR ALLFAST OR ARM_SIN_Q15) -target_sources(CMSISDSPFastMath PRIVATE arm_sin_q15.c) -endif() - -if (NOT CONFIGTABLE OR ALLFAST OR ARM_SIN_Q31) -target_sources(CMSISDSPFastMath PRIVATE arm_sin_q31.c) -endif() - -if (NOT CONFIGTABLE OR ALLFAST OR ARM_SQRT_Q31) -target_sources(CMSISDSPFastMath PRIVATE arm_sqrt_q31.c) -endif() - -if (NOT CONFIGTABLE OR ALLFAST OR ARM_SQRT_Q15) -target_sources(CMSISDSPFastMath PRIVATE arm_sqrt_q15.c) -endif() - -target_sources(CMSISDSPFastMath PRIVATE arm_vlog_f32.c) -target_sources(CMSISDSPFastMath PRIVATE arm_vlog_f64.c) -target_sources(CMSISDSPFastMath PRIVATE arm_vexp_f32.c) -target_sources(CMSISDSPFastMath PRIVATE arm_vexp_f64.c) - -target_sources(CMSISDSPFastMath PRIVATE arm_vlog_q31.c) -target_sources(CMSISDSPFastMath PRIVATE arm_vlog_q15.c) - - -if ((NOT ARMAC5) AND (NOT DISABLEFLOAT16)) -target_sources(CMSISDSPFastMath PRIVATE arm_vlog_f16.c) -target_sources(CMSISDSPFastMath PRIVATE arm_vexp_f16.c) -target_sources(CMSISDSPFastMath PRIVATE arm_vinverse_f16.c) -target_sources(CMSISDSPFastMath PRIVATE arm_atan2_f16.c) -endif() - -target_sources(CMSISDSPFastMath PRIVATE arm_divide_q15.c) -target_sources(CMSISDSPFastMath PRIVATE arm_divide_q31.c) - -target_sources(CMSISDSPFastMath PRIVATE arm_atan2_f32.c) -target_sources(CMSISDSPFastMath PRIVATE arm_atan2_q31.c) -target_sources(CMSISDSPFastMath PRIVATE arm_atan2_q15.c) - - -### Includes -target_include_directories(CMSISDSPFastMath PUBLIC "${DSP}/Include") - - - +cmake_minimum_required (VERSION 3.6) + +project(CMSISDSPFastMath) + + +file(GLOB SRC "./*_*.c") + +add_library(CMSISDSPFastMath STATIC) + +include(interpol) +interpol(CMSISDSPFastMath) + +if (CONFIGTABLE AND ALLFAST) + target_compile_definitions(CMSISDSPFastMath PUBLIC ARM_ALL_FAST_TABLES) +endif() + +if (NOT CONFIGTABLE OR ALLFAST OR ARM_COS_F32) +target_sources(CMSISDSPFastMath PRIVATE arm_cos_f32.c) +endif() + +if (NOT CONFIGTABLE OR ALLFAST OR ARM_COS_Q15) +target_sources(CMSISDSPFastMath PRIVATE arm_cos_q15.c) +endif() + +if (NOT CONFIGTABLE OR ALLFAST OR ARM_COS_Q31) +target_sources(CMSISDSPFastMath PRIVATE arm_cos_q31.c) +endif() + +if (NOT CONFIGTABLE OR ALLFAST OR ARM_SIN_F32) +target_sources(CMSISDSPFastMath PRIVATE arm_sin_f32.c) +endif() + +if (NOT CONFIGTABLE OR ALLFAST OR ARM_SIN_Q15) +target_sources(CMSISDSPFastMath PRIVATE arm_sin_q15.c) +endif() + +if (NOT CONFIGTABLE OR ALLFAST OR ARM_SIN_Q31) +target_sources(CMSISDSPFastMath PRIVATE arm_sin_q31.c) +endif() + +target_sources(CMSISDSPFastMath PRIVATE arm_sqrt_q15.c) +target_sources(CMSISDSPFastMath PRIVATE arm_sqrt_q31.c) + + +configdsp(CMSISDSPFastMath ..) + +### Includes +target_include_directories(CMSISDSPFastMath PUBLIC "${DSP}/../../Include") + + + diff --git a/Drivers/CMSIS/DSP/Source/FastMathFunctions/FastMathFunctions.c b/Drivers/CMSIS/DSP/Source/FastMathFunctions/FastMathFunctions.c index 0146d61..a856678 100644 --- a/Drivers/CMSIS/DSP/Source/FastMathFunctions/FastMathFunctions.c +++ b/Drivers/CMSIS/DSP/Source/FastMathFunctions/FastMathFunctions.c @@ -1,75 +1,37 @@ -/* ---------------------------------------------------------------------- - * Project: CMSIS DSP Library - * Title: FastMathFunctions.c - * Description: Combination of all fast math function source files. - * - * $Date: 16. March 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. - */ - -#if !defined(ARM_DSP_CONFIG_TABLES) || defined(ARM_FAST_ALLOW_TABLES) - -#if !defined(ARM_DSP_CONFIG_TABLES) || defined(ARM_ALL_FAST_TABLES) || defined(ARM_TABLE_SIN_F32) -#include "arm_cos_f32.c" -#endif - -#if !defined(ARM_DSP_CONFIG_TABLES) || defined(ARM_ALL_FAST_TABLES) || defined(ARM_TABLE_SIN_Q15) -#include "arm_cos_q15.c" -#endif - -#if !defined(ARM_DSP_CONFIG_TABLES) || defined(ARM_ALL_FAST_TABLES) || defined(ARM_TABLE_SIN_Q31) -#include "arm_cos_q31.c" -#endif - -#if !defined(ARM_DSP_CONFIG_TABLES) || defined(ARM_ALL_FAST_TABLES) || defined(ARM_TABLE_SIN_F32) -#include "arm_sin_f32.c" -#endif - -#if !defined(ARM_DSP_CONFIG_TABLES) || defined(ARM_ALL_FAST_TABLES) || defined(ARM_TABLE_SIN_Q15) -#include "arm_sin_q15.c" -#endif - -#if !defined(ARM_DSP_CONFIG_TABLES) || defined(ARM_ALL_FAST_TABLES) || defined(ARM_TABLE_SIN_Q31) -#include "arm_sin_q31.c" -#endif - -#if !defined(ARM_DSP_CONFIG_TABLES) || defined(ARM_ALL_FAST_TABLES) || defined(ARM_TABLE_SQRT_Q31) -#include "arm_sqrt_q31.c" -#endif - -#if !defined(ARM_DSP_CONFIG_TABLES) || defined(ARM_ALL_FAST_TABLES) || defined(ARM_TABLE_SQRT_Q15) -#include "arm_sqrt_q15.c" -#endif - -#endif - -#include "arm_vexp_f32.c" -#include "arm_vexp_f64.c" -#include "arm_vlog_f32.c" -#include "arm_vlog_f64.c" -#include "arm_divide_q15.c" -#include "arm_divide_q31.c" -#include "arm_vlog_q31.c" -#include "arm_vlog_q15.c" -#include "arm_atan2_f32.c" -#include "arm_atan2_q31.c" -#include "arm_atan2_q15.c" +/* ---------------------------------------------------------------------- + * Project: CMSIS DSP Library + * Title: FastMathFunctions.c + * Description: Combination of all fast math function source files. + * + * $Date: 18. March 2019 + * $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_cos_f32.c" +#include "arm_cos_q15.c" +#include "arm_cos_q31.c" +#include "arm_sin_f32.c" +#include "arm_sin_q15.c" +#include "arm_sin_q31.c" +#include "arm_sqrt_q15.c" +#include "arm_sqrt_q31.c" + diff --git a/Drivers/CMSIS/DSP/Source/FastMathFunctions/arm_cos_f32.c b/Drivers/CMSIS/DSP/Source/FastMathFunctions/arm_cos_f32.c index de037c8..78207d8 100644 --- a/Drivers/CMSIS/DSP/Source/FastMathFunctions/arm_cos_f32.c +++ b/Drivers/CMSIS/DSP/Source/FastMathFunctions/arm_cos_f32.c @@ -1,121 +1,122 @@ -/* ---------------------------------------------------------------------- - * Project: CMSIS DSP Library - * Title: arm_cos_f32.c - * Description: Fast cosine calculation for floating-point values - * - * $Date: 23 April 2021 - * $Revision: V1.9.0 - * - * Target Processor: Cortex-M and Cortex-A cores - * -------------------------------------------------------------------- */ -/* - * Copyright (C) 2010-2021 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 "dsp/fast_math_functions.h" -#include "arm_common_tables.h" - -/** - @ingroup groupFastMath - */ - -/** - @defgroup cos Cosine - - Computes the trigonometric cosine function using a combination of table lookup - and linear interpolation. There are separate functions for - Q15, Q31, and floating-point data types. - The input to the floating-point version is in radians while the - fixed-point Q15 and Q31 have a scaled input with the range - [0 +0.9999] mapping to [0 2*pi). The fixed-point range is chosen so that a - value of 2*pi wraps around to 0. - - The implementation is based on table lookup using 512 values together with linear interpolation. - The steps used are: - -# Calculation of the nearest integer table index - -# Compute the fractional portion (fract) of the table index. - -# The final result equals (1.0f-fract)*a + fract*b; - - where - 
-     a = Table[index];
-     b = Table[index+1];
-
- */ - -/** - @addtogroup cos - @{ - */ - -/** - @brief Fast approximation to the trigonometric cosine function for floating-point data. - @param[in] x input value in radians - @return cos(x) - */ -float32_t arm_cos_f32( - float32_t x) -{ - float32_t cosVal, fract, in; /* Temporary input, output variables */ - uint16_t index; /* Index variable */ - float32_t a, b; /* Two nearest output values */ - int32_t n; - float32_t findex; - - /* input x is in radians */ - /* Scale input to [0 1] range from [0 2*PI] , divide input by 2*pi, add 0.25 (pi/2) to read sine table */ - in = x * 0.159154943092f + 0.25f; - - /* Calculation of floor value of input */ - n = (int32_t) in; - - /* Make negative values towards -infinity */ - if (in < 0.0f) - { - n--; - } - - /* Map input value to [0 1] */ - in = in - (float32_t) n; - - /* Calculation of index of the table */ - findex = (float32_t)FAST_MATH_TABLE_SIZE * in; - index = (uint16_t)findex; - - /* when "in" is exactly 1, we need to rotate the index down to 0 */ - if (index >= FAST_MATH_TABLE_SIZE) { - index = 0; - findex -= (float32_t)FAST_MATH_TABLE_SIZE; - } - - /* fractional value calculation */ - fract = findex - (float32_t) index; - - /* Read two nearest values of input value from the cos table */ - a = sinTable_f32[index]; - b = sinTable_f32[index+1]; - - /* Linear interpolation process */ - cosVal = (1.0f - fract) * a + fract * b; - - /* Return output value */ - return (cosVal); -} - -/** - @} end of cos group - */ +/* ---------------------------------------------------------------------- + * Project: CMSIS DSP Library + * Title: arm_cos_f32.c + * Description: Fast cosine calculation for floating-point values + * + * $Date: 18. March 2019 + * $Revision: V1.6.0 + * + * Target Processor: Cortex-M cores + * -------------------------------------------------------------------- */ +/* + * Copyright (C) 2010-2019 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_math.h" +#include "arm_common_tables.h" + +/** + @ingroup groupFastMath + */ + +/** + @defgroup cos Cosine + + Computes the trigonometric cosine function using a combination of table lookup + and linear interpolation. There are separate functions for + Q15, Q31, and floating-point data types. + The input to the floating-point version is in radians while the + fixed-point Q15 and Q31 have a scaled input with the range + [0 +0.9999] mapping to [0 2*pi). The fixed-point range is chosen so that a + value of 2*pi wraps around to 0. + + The implementation is based on table lookup using 256 values together with linear interpolation. + The steps used are: + -# Calculation of the nearest integer table index + -# Compute the fractional portion (fract) of the table index. + -# The final result equals (1.0f-fract)*a + fract*b; + + where + 
+     b = Table[index];
+     c = Table[index+1];
+
+ */ + +/** + @addtogroup cos + @{ + */ + +/** + @brief Fast approximation to the trigonometric cosine function for floating-point data. + @param[in] x input value in radians + @return cos(x) + */ + +float32_t arm_cos_f32( + float32_t x) +{ + float32_t cosVal, fract, in; /* Temporary input, output variables */ + uint16_t index; /* Index variable */ + float32_t a, b; /* Two nearest output values */ + int32_t n; + float32_t findex; + + /* input x is in radians */ + /* Scale input to [0 1] range from [0 2*PI] , divide input by 2*pi, add 0.25 (pi/2) to read sine table */ + in = x * 0.159154943092f + 0.25f; + + /* Calculation of floor value of input */ + n = (int32_t) in; + + /* Make negative values towards -infinity */ + if (in < 0.0f) + { + n--; + } + + /* Map input value to [0 1] */ + in = in - (float32_t) n; + + /* Calculation of index of the table */ + findex = (float32_t)FAST_MATH_TABLE_SIZE * in; + index = (uint16_t)findex; + + /* when "in" is exactly 1, we need to rotate the index down to 0 */ + if (index >= FAST_MATH_TABLE_SIZE) { + index = 0; + findex -= (float32_t)FAST_MATH_TABLE_SIZE; + } + + /* fractional value calculation */ + fract = findex - (float32_t) index; + + /* Read two nearest values of input value from the cos table */ + a = sinTable_f32[index]; + b = sinTable_f32[index+1]; + + /* Linear interpolation process */ + cosVal = (1.0f - fract) * a + fract * b; + + /* Return output value */ + return (cosVal); +} + +/** + @} end of cos group + */ diff --git a/Drivers/CMSIS/DSP/Source/FastMathFunctions/arm_cos_q15.c b/Drivers/CMSIS/DSP/Source/FastMathFunctions/arm_cos_q15.c index 6b552f1..aef47da 100644 --- a/Drivers/CMSIS/DSP/Source/FastMathFunctions/arm_cos_q15.c +++ b/Drivers/CMSIS/DSP/Source/FastMathFunctions/arm_cos_q15.c @@ -1,84 +1,84 @@ -/* ---------------------------------------------------------------------- - * Project: CMSIS DSP Library - * Title: arm_cos_q15.c - * Description: Fast cosine calculation for Q15 values - * - * $Date: 23 April 2021 - * $Revision: V1.9.0 - * - * Target Processor: Cortex-M and Cortex-A cores - * -------------------------------------------------------------------- */ -/* - * Copyright (C) 2010-2021 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 "dsp/fast_math_functions.h" -#include "arm_common_tables.h" - -/** - @ingroup groupFastMath - */ - -/** - @addtogroup cos - @{ - */ - -/** - @brief Fast approximation to the trigonometric cosine function for Q15 data. - @param[in] x Scaled input value in radians - @return cos(x) - - The Q15 input value is in the range [0 +0.9999] and is mapped to a radian value in the range [0 2*PI). - */ - -q15_t arm_cos_q15( - q15_t x) -{ - q15_t cosVal; /* Temporary input, output variables */ - int32_t index; /* Index variable */ - q15_t a, b; /* Two nearest output values */ - q15_t fract; /* Temporary values for fractional values */ - - /* add 0.25 (pi/2) to read sine table */ - x = (uint16_t)x + 0x2000; - if (x < 0) - { /* convert negative numbers to corresponding positive ones */ - x = (uint16_t)x + 0x8000; - } - - /* Calculate the nearest index */ - index = (uint32_t)x >> FAST_MATH_Q15_SHIFT; - - /* Calculation of fractional value */ - fract = (x - (index << FAST_MATH_Q15_SHIFT)) << 9; - - /* Read two nearest values of input value from the sin table */ - a = sinTable_q15[index]; - b = sinTable_q15[index+1]; - - /* Linear interpolation process */ - cosVal = (q31_t) (0x8000 - fract) * a >> 16; - cosVal = (q15_t) ((((q31_t) cosVal << 16) + ((q31_t) fract * b)) >> 16); - - /* Return output value */ - return (cosVal << 1); -} - -/** - @} end of cos group - */ +/* ---------------------------------------------------------------------- + * Project: CMSIS DSP Library + * Title: arm_cos_q15.c + * Description: Fast cosine calculation for Q15 values + * + * $Date: 18. March 2019 + * $Revision: V1.6.0 + * + * Target Processor: Cortex-M cores + * -------------------------------------------------------------------- */ +/* + * Copyright (C) 2010-2019 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_math.h" +#include "arm_common_tables.h" + +/** + @ingroup groupFastMath + */ + +/** + @addtogroup cos + @{ + */ + +/** + @brief Fast approximation to the trigonometric cosine function for Q15 data. + @param[in] x Scaled input value in radians + @return cos(x) + + The Q15 input value is in the range [0 +0.9999] and is mapped to a radian value in the range [0 2*PI). + */ + +q15_t arm_cos_q15( + q15_t x) +{ + q15_t cosVal; /* Temporary input, output variables */ + int32_t index; /* Index variable */ + q15_t a, b; /* Two nearest output values */ + q15_t fract; /* Temporary values for fractional values */ + + /* add 0.25 (pi/2) to read sine table */ + x = (uint16_t)x + 0x2000; + if (x < 0) + { /* convert negative numbers to corresponding positive ones */ + x = (uint16_t)x + 0x8000; + } + + /* Calculate the nearest index */ + index = (uint32_t)x >> FAST_MATH_Q15_SHIFT; + + /* Calculation of fractional value */ + fract = (x - (index << FAST_MATH_Q15_SHIFT)) << 9; + + /* Read two nearest values of input value from the sin table */ + a = sinTable_q15[index]; + b = sinTable_q15[index+1]; + + /* Linear interpolation process */ + cosVal = (q31_t) (0x8000 - fract) * a >> 16; + cosVal = (q15_t) ((((q31_t) cosVal << 16) + ((q31_t) fract * b)) >> 16); + + /* Return output value */ + return (cosVal << 1); +} + +/** + @} end of cos group + */ diff --git a/Drivers/CMSIS/DSP/Source/FastMathFunctions/arm_cos_q31.c b/Drivers/CMSIS/DSP/Source/FastMathFunctions/arm_cos_q31.c index 8386c8f..fa42eee 100644 --- a/Drivers/CMSIS/DSP/Source/FastMathFunctions/arm_cos_q31.c +++ b/Drivers/CMSIS/DSP/Source/FastMathFunctions/arm_cos_q31.c @@ -1,84 +1,84 @@ -/* ---------------------------------------------------------------------- - * Project: CMSIS DSP Library - * Title: arm_cos_q31.c - * Description: Fast cosine calculation for Q31 values - * - * $Date: 23 April 2021 - * $Revision: V1.9.0 - * - * Target Processor: Cortex-M and Cortex-A cores - * -------------------------------------------------------------------- */ -/* - * Copyright (C) 2010-2021 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 "dsp/fast_math_functions.h" -#include "arm_common_tables.h" - -/** - @ingroup groupFastMath - */ - -/** - @addtogroup cos - @{ - */ - -/** - @brief Fast approximation to the trigonometric cosine function for Q31 data. - @param[in] x Scaled input value in radians - @return cos(x) - - The Q31 input value is in the range [0 +0.9999] and is mapped to a radian value in the range [0 2*PI). - */ - -q31_t arm_cos_q31( - q31_t x) -{ - q31_t cosVal; /* Temporary input, output variables */ - int32_t index; /* Index variable */ - q31_t a, b; /* Two nearest output values */ - q31_t fract; /* Temporary values for fractional values */ - - /* add 0.25 (pi/2) to read sine table */ - x = (uint32_t)x + 0x20000000; - if (x < 0) - { /* convert negative numbers to corresponding positive ones */ - x = (uint32_t)x + 0x80000000; - } - - /* Calculate the nearest index */ - index = (uint32_t)x >> FAST_MATH_Q31_SHIFT; - - /* Calculation of fractional value */ - fract = (x - (index << FAST_MATH_Q31_SHIFT)) << 9; - - /* Read two nearest values of input value from the sin table */ - a = sinTable_q31[index]; - b = sinTable_q31[index+1]; - - /* Linear interpolation process */ - cosVal = (q63_t) (0x80000000 - fract) * a >> 32; - cosVal = (q31_t) ((((q63_t) cosVal << 32) + ((q63_t) fract * b)) >> 32); - - /* Return output value */ - return (cosVal << 1); -} - -/** - @} end of cos group - */ +/* ---------------------------------------------------------------------- + * Project: CMSIS DSP Library + * Title: arm_cos_q31.c + * Description: Fast cosine calculation for Q31 values + * + * $Date: 18. March 2019 + * $Revision: V1.6.0 + * + * Target Processor: Cortex-M cores + * -------------------------------------------------------------------- */ +/* + * Copyright (C) 2010-2019 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_math.h" +#include "arm_common_tables.h" + +/** + @ingroup groupFastMath + */ + +/** + @addtogroup cos + @{ + */ + +/** + @brief Fast approximation to the trigonometric cosine function for Q31 data. + @param[in] x Scaled input value in radians + @return cos(x) + + The Q31 input value is in the range [0 +0.9999] and is mapped to a radian value in the range [0 2*PI). + */ + +q31_t arm_cos_q31( + q31_t x) +{ + q31_t cosVal; /* Temporary input, output variables */ + int32_t index; /* Index variable */ + q31_t a, b; /* Two nearest output values */ + q31_t fract; /* Temporary values for fractional values */ + + /* add 0.25 (pi/2) to read sine table */ + x = (uint32_t)x + 0x20000000; + if (x < 0) + { /* convert negative numbers to corresponding positive ones */ + x = (uint32_t)x + 0x80000000; + } + + /* Calculate the nearest index */ + index = (uint32_t)x >> FAST_MATH_Q31_SHIFT; + + /* Calculation of fractional value */ + fract = (x - (index << FAST_MATH_Q31_SHIFT)) << 9; + + /* Read two nearest values of input value from the sin table */ + a = sinTable_q31[index]; + b = sinTable_q31[index+1]; + + /* Linear interpolation process */ + cosVal = (q63_t) (0x80000000 - fract) * a >> 32; + cosVal = (q31_t) ((((q63_t) cosVal << 32) + ((q63_t) fract * b)) >> 32); + + /* Return output value */ + return (cosVal << 1); +} + +/** + @} end of cos group + */ diff --git a/Drivers/CMSIS/DSP/Source/FastMathFunctions/arm_sin_f32.c b/Drivers/CMSIS/DSP/Source/FastMathFunctions/arm_sin_f32.c index 2fcd6bb..674765e 100644 --- a/Drivers/CMSIS/DSP/Source/FastMathFunctions/arm_sin_f32.c +++ b/Drivers/CMSIS/DSP/Source/FastMathFunctions/arm_sin_f32.c @@ -1,122 +1,122 @@ -/* ---------------------------------------------------------------------- - * Project: CMSIS DSP Library - * Title: arm_sin_f32.c - * Description: Fast sine calculation for floating-point values - * - * $Date: 23 April 2021 - * $Revision: V1.9.0 - * - * Target Processor: Cortex-M and Cortex-A cores - * -------------------------------------------------------------------- */ -/* - * Copyright (C) 2010-2021 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 "dsp/fast_math_functions.h" -#include "arm_common_tables.h" - -/** - @ingroup groupFastMath - */ - -/** - @defgroup sin Sine - - Computes the trigonometric sine function using a combination of table lookup - and linear interpolation. There are separate functions for - Q15, Q31, and floating-point data types. - The input to the floating-point version is in radians while the - fixed-point Q15 and Q31 have a scaled input with the range - [0 +0.9999] mapping to [0 2*pi). The fixed-point range is chosen so that a - value of 2*pi wraps around to 0. - - The implementation is based on table lookup using 512 values together with linear interpolation. - The steps used are: - -# Calculation of the nearest integer table index - -# Compute the fractional portion (fract) of the table index. - -# The final result equals (1.0f-fract)*a + fract*b; - - where - 
-     b = Table[index];
-     c = Table[index+1];
-
- */ - -/** - @addtogroup sin - @{ - */ - -/** - @brief Fast approximation to the trigonometric sine function for floating-point data. - @param[in] x input value in radians. - @return sin(x) - */ - -float32_t arm_sin_f32( - float32_t x) -{ - float32_t sinVal, fract, in; /* Temporary input, output variables */ - uint16_t index; /* Index variable */ - float32_t a, b; /* Two nearest output values */ - int32_t n; - float32_t findex; - - /* input x is in radians */ - /* Scale input to [0 1] range from [0 2*PI] , divide input by 2*pi */ - in = x * 0.159154943092f; - - /* Calculation of floor value of input */ - n = (int32_t) in; - - /* Make negative values towards -infinity */ - if (in < 0.0f) - { - n--; - } - - /* Map input value to [0 1] */ - in = in - (float32_t) n; - - /* Calculation of index of the table */ - findex = (float32_t)FAST_MATH_TABLE_SIZE * in; - index = (uint16_t)findex; - - /* when "in" is exactly 1, we need to rotate the index down to 0 */ - if (index >= FAST_MATH_TABLE_SIZE) { - index = 0; - findex -= (float32_t)FAST_MATH_TABLE_SIZE; - } - - /* fractional value calculation */ - fract = findex - (float32_t) index; - - /* Read two nearest values of input value from the sin table */ - a = sinTable_f32[index]; - b = sinTable_f32[index+1]; - - /* Linear interpolation process */ - sinVal = (1.0f - fract) * a + fract * b; - - /* Return output value */ - return (sinVal); -} - -/** - @} end of sin group - */ +/* ---------------------------------------------------------------------- + * Project: CMSIS DSP Library + * Title: arm_sin_f32.c + * Description: Fast sine calculation for floating-point values + * + * $Date: 18. March 2019 + * $Revision: V1.6.0 + * + * Target Processor: Cortex-M cores + * -------------------------------------------------------------------- */ +/* + * Copyright (C) 2010-2019 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_math.h" +#include "arm_common_tables.h" + +/** + @ingroup groupFastMath + */ + +/** + @defgroup sin Sine + + Computes the trigonometric sine function using a combination of table lookup + and linear interpolation. There are separate functions for + Q15, Q31, and floating-point data types. + The input to the floating-point version is in radians while the + fixed-point Q15 and Q31 have a scaled input with the range + [0 +0.9999] mapping to [0 2*pi). The fixed-point range is chosen so that a + value of 2*pi wraps around to 0. + + The implementation is based on table lookup using 256 values together with linear interpolation. + The steps used are: + -# Calculation of the nearest integer table index + -# Compute the fractional portion (fract) of the table index. + -# The final result equals (1.0f-fract)*a + fract*b; + + where + 
+     b = Table[index];
+     c = Table[index+1];
+
+ */ + +/** + @addtogroup sin + @{ + */ + +/** + @brief Fast approximation to the trigonometric sine function for floating-point data. + @param[in] x input value in radians. + @return sin(x) + */ + +float32_t arm_sin_f32( + float32_t x) +{ + float32_t sinVal, fract, in; /* Temporary input, output variables */ + uint16_t index; /* Index variable */ + float32_t a, b; /* Two nearest output values */ + int32_t n; + float32_t findex; + + /* input x is in radians */ + /* Scale input to [0 1] range from [0 2*PI] , divide input by 2*pi */ + in = x * 0.159154943092f; + + /* Calculation of floor value of input */ + n = (int32_t) in; + + /* Make negative values towards -infinity */ + if (in < 0.0f) + { + n--; + } + + /* Map input value to [0 1] */ + in = in - (float32_t) n; + + /* Calculation of index of the table */ + findex = (float32_t)FAST_MATH_TABLE_SIZE * in; + index = (uint16_t)findex; + + /* when "in" is exactly 1, we need to rotate the index down to 0 */ + if (index >= FAST_MATH_TABLE_SIZE) { + index = 0; + findex -= (float32_t)FAST_MATH_TABLE_SIZE; + } + + /* fractional value calculation */ + fract = findex - (float32_t) index; + + /* Read two nearest values of input value from the sin table */ + a = sinTable_f32[index]; + b = sinTable_f32[index+1]; + + /* Linear interpolation process */ + sinVal = (1.0f - fract) * a + fract * b; + + /* Return output value */ + return (sinVal); +} + +/** + @} end of sin group + */ diff --git a/Drivers/CMSIS/DSP/Source/FastMathFunctions/arm_sin_q15.c b/Drivers/CMSIS/DSP/Source/FastMathFunctions/arm_sin_q15.c index 81f0d62..edc8b38 100644 --- a/Drivers/CMSIS/DSP/Source/FastMathFunctions/arm_sin_q15.c +++ b/Drivers/CMSIS/DSP/Source/FastMathFunctions/arm_sin_q15.c @@ -1,83 +1,77 @@ -/* ---------------------------------------------------------------------- - * Project: CMSIS DSP Library - * Title: arm_sin_q15.c - * Description: Fast sine calculation for Q15 values - * - * $Date: 23 April 2021 - * $Revision: V1.9.0 - * - * Target Processor: Cortex-M and Cortex-A cores - * -------------------------------------------------------------------- */ -/* - * Copyright (C) 2010-2021 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 "dsp/fast_math_functions.h" -#include "arm_common_tables.h" - -/** - @ingroup groupFastMath - */ - -/** - @addtogroup sin - @{ - */ - -/** - @brief Fast approximation to the trigonometric sine function for Q15 data. - @param[in] x Scaled input value in radians - @return sin(x) - - The Q15 input value is in the range [0 +0.9999] and is mapped to a radian value in the range [0 2*PI). - */ - -q15_t arm_sin_q15( - q15_t x) -{ - q15_t sinVal; /* Temporary input, output variables */ - int32_t index; /* Index variable */ - q15_t a, b; /* Two nearest output values */ - q15_t fract; /* Temporary values for fractional values */ - - - if (x < 0) - { /* convert negative numbers to corresponding positive ones */ - x = (uint16_t)x + 0x8000; - } - - /* Calculate the nearest index */ - index = (uint32_t)x >> FAST_MATH_Q15_SHIFT; - - /* Calculation of fractional value */ - fract = (x - (index << FAST_MATH_Q15_SHIFT)) << 9; - - /* Read two nearest values of input value from the sin table */ - a = sinTable_q15[index]; - b = sinTable_q15[index+1]; - - /* Linear interpolation process */ - sinVal = (q31_t) (0x8000 - fract) * a >> 16; - sinVal = (q15_t) ((((q31_t) sinVal << 16) + ((q31_t) fract * b)) >> 16); - - /* Return output value */ - return (sinVal << 1); -} - -/** - @} end of sin group - */ +/* ---------------------------------------------------------------------- + * Project: CMSIS DSP Library + * Title: arm_sin_q15.c + * Description: Fast sine calculation for Q15 values + * + * $Date: 18. March 2019 + * $Revision: V1.6.0 + * + * Target Processor: Cortex-M cores + * -------------------------------------------------------------------- */ +/* + * Copyright (C) 2010-2019 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_math.h" +#include "arm_common_tables.h" + +/** + @ingroup groupFastMath + */ + +/** + @addtogroup sin + @{ + */ + +/** + @brief Fast approximation to the trigonometric sine function for Q15 data. + @param[in] x Scaled input value in radians + @return sin(x) + + The Q15 input value is in the range [0 +0.9999] and is mapped to a radian value in the range [0 2*PI). + */ + +q15_t arm_sin_q15( + q15_t x) +{ + q15_t sinVal; /* Temporary input, output variables */ + int32_t index; /* Index variable */ + q15_t a, b; /* Two nearest output values */ + q15_t fract; /* Temporary values for fractional values */ + + /* Calculate the nearest index */ + index = (uint32_t)x >> FAST_MATH_Q15_SHIFT; + + /* Calculation of fractional value */ + fract = (x - (index << FAST_MATH_Q15_SHIFT)) << 9; + + /* Read two nearest values of input value from the sin table */ + a = sinTable_q15[index]; + b = sinTable_q15[index+1]; + + /* Linear interpolation process */ + sinVal = (q31_t) (0x8000 - fract) * a >> 16; + sinVal = (q15_t) ((((q31_t) sinVal << 16) + ((q31_t) fract * b)) >> 16); + + /* Return output value */ + return (sinVal << 1); +} + +/** + @} end of sin group + */ diff --git a/Drivers/CMSIS/DSP/Source/FastMathFunctions/arm_sin_q31.c b/Drivers/CMSIS/DSP/Source/FastMathFunctions/arm_sin_q31.c index 06262fd..849cc11 100644 --- a/Drivers/CMSIS/DSP/Source/FastMathFunctions/arm_sin_q31.c +++ b/Drivers/CMSIS/DSP/Source/FastMathFunctions/arm_sin_q31.c @@ -1,82 +1,77 @@ -/* ---------------------------------------------------------------------- - * Project: CMSIS DSP Library - * Title: arm_sin_q31.c - * Description: Fast sine calculation for Q31 values - * - * $Date: 23 April 2021 - * $Revision: V1.9.0 - * - * Target Processor: Cortex-M and Cortex-A cores - * -------------------------------------------------------------------- */ -/* - * Copyright (C) 2010-2021 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 "dsp/fast_math_functions.h" -#include "arm_common_tables.h" - -/** - @ingroup groupFastMath - */ - -/** - @addtogroup sin - @{ - */ - -/** - @brief Fast approximation to the trigonometric sine function for Q31 data. - @param[in] x Scaled input value in radians - @return sin(x) - - The Q31 input value is in the range [0 +0.9999] and is mapped to a radian value in the range [0 2*PI). - */ - -q31_t arm_sin_q31( - q31_t x) -{ - q31_t sinVal; /* Temporary variables for input, output */ - int32_t index; /* Index variable */ - q31_t a, b; /* Two nearest output values */ - q31_t fract; /* Temporary values for fractional values */ - - if (x < 0) - { /* convert negative numbers to corresponding positive ones */ - x = (uint32_t)x + 0x80000000; - } - - /* Calculate the nearest index */ - index = (uint32_t)x >> FAST_MATH_Q31_SHIFT; - - /* Calculation of fractional value */ - fract = (x - (index << FAST_MATH_Q31_SHIFT)) << 9; - - /* Read two nearest values of input value from the sin table */ - a = sinTable_q31[index]; - b = sinTable_q31[index+1]; - - /* Linear interpolation process */ - sinVal = (q63_t) (0x80000000 - fract) * a >> 32; - sinVal = (q31_t) ((((q63_t) sinVal << 32) + ((q63_t) fract * b)) >> 32); - - /* Return output value */ - return (sinVal << 1); -} - -/** - @} end of sin group - */ +/* ---------------------------------------------------------------------- + * Project: CMSIS DSP Library + * Title: arm_sin_q31.c + * Description: Fast sine calculation for Q31 values + * + * $Date: 18. March 2019 + * $Revision: V1.6.0 + * + * Target Processor: Cortex-M cores + * -------------------------------------------------------------------- */ +/* + * Copyright (C) 2010-2019 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_math.h" +#include "arm_common_tables.h" + +/** + @ingroup groupFastMath + */ + +/** + @addtogroup sin + @{ + */ + +/** + @brief Fast approximation to the trigonometric sine function for Q31 data. + @param[in] x Scaled input value in radians + @return sin(x) + + The Q31 input value is in the range [0 +0.9999] and is mapped to a radian value in the range [0 2*PI). + */ + +q31_t arm_sin_q31( + q31_t x) +{ + q31_t sinVal; /* Temporary variables for input, output */ + int32_t index; /* Index variable */ + q31_t a, b; /* Two nearest output values */ + q31_t fract; /* Temporary values for fractional values */ + + /* Calculate the nearest index */ + index = (uint32_t)x >> FAST_MATH_Q31_SHIFT; + + /* Calculation of fractional value */ + fract = (x - (index << FAST_MATH_Q31_SHIFT)) << 9; + + /* Read two nearest values of input value from the sin table */ + a = sinTable_q31[index]; + b = sinTable_q31[index+1]; + + /* Linear interpolation process */ + sinVal = (q63_t) (0x80000000 - fract) * a >> 32; + sinVal = (q31_t) ((((q63_t) sinVal << 32) + ((q63_t) fract * b)) >> 32); + + /* Return output value */ + return (sinVal << 1); +} + +/** + @} end of sin group + */ diff --git a/Drivers/CMSIS/DSP/Source/FastMathFunctions/arm_sqrt_q15.c b/Drivers/CMSIS/DSP/Source/FastMathFunctions/arm_sqrt_q15.c index defeebe..1dc7bc4 100644 --- a/Drivers/CMSIS/DSP/Source/FastMathFunctions/arm_sqrt_q15.c +++ b/Drivers/CMSIS/DSP/Source/FastMathFunctions/arm_sqrt_q15.c @@ -1,123 +1,144 @@ -/* ---------------------------------------------------------------------- - * Project: CMSIS DSP Library - * Title: arm_sqrt_q15.c - * Description: Q15 square root function - * - * $Date: 23 April 2021 - * $Revision: V1.9.0 - * - * Target Processor: Cortex-M and Cortex-A cores - * -------------------------------------------------------------------- */ -/* - * Copyright (C) 2010-2021 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 "dsp/fast_math_functions.h" -#include "arm_common_tables.h" - -/** - @ingroup groupFastMath - */ - -/** - @addtogroup SQRT - @{ - */ - -/** - @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 ARM_MATH_SUCCESS : input value is positive - - \ref ARM_MATH_ARGUMENT_ERROR : input value is negative; *pOut is set to 0 - */ - -#define Q12QUARTER 0x2000 -arm_status arm_sqrt_q15( - q15_t in, - q15_t * pOut) -{ - q15_t number, var1, signBits1,temp; - - number = in; - - /* If the input is a positive number then compute the signBits. */ - if (number > 0) - { - signBits1 = __CLZ(number) - 17; - - /* Shift by the number of signBits1 */ - if ((signBits1 % 2) == 0) - { - number = number << signBits1; - } - else - { - number = number << (signBits1 - 1); - } - /* Start value for 1/sqrt(x) for the Newton iteration */ - var1 = sqrt_initial_lut_q15[(number>> 11) - (Q12QUARTER >> 11)]; - - /* 0.5 var1 * (3 - number * var1 * var1) */ - /* 1st iteration */ - - temp = ((q31_t) var1 * var1) >> 12; - temp = ((q31_t) number * temp) >> 15; - temp = 0x3000 - temp; - var1 = ((q31_t) var1 * temp) >> 13; - - temp = ((q31_t) var1 * var1) >> 12; - temp = ((q31_t) number * temp) >> 15; - temp = 0x3000 - temp; - var1 = ((q31_t) var1 * temp) >> 13; - - temp = ((q31_t) var1 * var1) >> 12; - temp = ((q31_t) number * temp) >> 15; - temp = 0x3000 - temp; - var1 = ((q31_t) var1 * temp) >> 13; - - /* Multiply the inverse square root with the original value */ - - var1 = ((q15_t) (((q31_t) number * var1) >> 12)); - - /* Shift the output down accordingly */ - if ((signBits1 % 2) == 0) - { - var1 = var1 >> (signBits1 / 2); - } - else - { - var1 = var1 >> ((signBits1 - 1) / 2); - } - *pOut = var1; - - - return (ARM_MATH_SUCCESS); - } - /* If the number is a negative number then store zero as its square root value */ - else - { - *pOut = 0; - - return (ARM_MATH_ARGUMENT_ERROR); - } -} - -/** - @} end of SQRT group - */ +/* ---------------------------------------------------------------------- + * Project: CMSIS DSP Library + * Title: arm_sqrt_q15.c + * Description: Q15 square root function + * + * $Date: 18. March 2019 + * $Revision: V1.6.0 + * + * Target Processor: Cortex-M cores + * -------------------------------------------------------------------- */ +/* + * Copyright (C) 2010-2019 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_math.h" +#include "arm_common_tables.h" + +/** + @ingroup groupFastMath + */ + +/** + @addtogroup SQRT + @{ + */ + +/** + @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 ARM_MATH_SUCCESS : input value is positive + - \ref ARM_MATH_ARGUMENT_ERROR : input value is negative; *pOut is set to 0 + */ + +arm_status arm_sqrt_q15( + q15_t in, + q15_t * pOut) +{ + q31_t bits_val1; + q15_t number, temp1, var1, signBits1, half; + float32_t temp_float1; + union + { + q31_t fracval; + float32_t floatval; + } tempconv; + + number = in; + + /* If the input is a positive number then compute the signBits. */ + if (number > 0) + { + signBits1 = __CLZ(number) - 17; + + /* Shift by the number of signBits1 */ + if ((signBits1 % 2) == 0) + { + number = number << signBits1; + } + else + { + number = number << (signBits1 - 1); + } + + /* Calculate half value of the number */ + half = number >> 1; + /* Store the number for later use */ + temp1 = number; + + /* Convert to float */ + temp_float1 = number * 3.051757812500000e-005f; + /* Store as integer */ + tempconv.floatval = temp_float1; + bits_val1 = tempconv.fracval; + /* Subtract the shifted value from the magic number to give intial guess */ + bits_val1 = 0x5f3759df - (bits_val1 >> 1); /* gives initial guess */ + /* Store as float */ + tempconv.fracval = bits_val1; + temp_float1 = tempconv.floatval; + /* Convert to integer format */ + var1 = (q31_t) (temp_float1 * 16384); + + /* 1st iteration */ + var1 = ((q15_t) ((q31_t) var1 * (0x3000 - + ((q15_t) + ((((q15_t) + (((q31_t) var1 * var1) >> 15)) * + (q31_t) half) >> 15))) >> 15)) << 2; + /* 2nd iteration */ + var1 = ((q15_t) ((q31_t) var1 * (0x3000 - + ((q15_t) + ((((q15_t) + (((q31_t) var1 * var1) >> 15)) * + (q31_t) half) >> 15))) >> 15)) << 2; + /* 3rd iteration */ + var1 = ((q15_t) ((q31_t) var1 * (0x3000 - + ((q15_t) + ((((q15_t) + (((q31_t) var1 * var1) >> 15)) * + (q31_t) half) >> 15))) >> 15)) << 2; + + /* Multiply the inverse square root with the original value */ + var1 = ((q15_t) (((q31_t) temp1 * var1) >> 15)) << 1; + + /* Shift the output down accordingly */ + if ((signBits1 % 2) == 0) + { + var1 = var1 >> (signBits1 / 2); + } + else + { + var1 = var1 >> ((signBits1 - 1) / 2); + } + *pOut = var1; + + return (ARM_MATH_SUCCESS); + } + /* If the number is a negative number then store zero as its square root value */ + else + { + *pOut = 0; + + return (ARM_MATH_ARGUMENT_ERROR); + } +} + +/** + @} end of SQRT group + */ diff --git a/Drivers/CMSIS/DSP/Source/FastMathFunctions/arm_sqrt_q31.c b/Drivers/CMSIS/DSP/Source/FastMathFunctions/arm_sqrt_q31.c index 8747670..9f2a731 100644 --- a/Drivers/CMSIS/DSP/Source/FastMathFunctions/arm_sqrt_q31.c +++ b/Drivers/CMSIS/DSP/Source/FastMathFunctions/arm_sqrt_q31.c @@ -1,126 +1,144 @@ -/* ---------------------------------------------------------------------- - * Project: CMSIS DSP Library - * Title: arm_sqrt_q31.c - * Description: Q31 square root function - * - * $Date: 23 April 2021 - * $Revision: V1.9.0 - * - * Target Processor: Cortex-M and Cortex-A cores - * -------------------------------------------------------------------- */ -/* - * Copyright (C) 2010-2021 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 "dsp/fast_math_functions.h" -#include "arm_common_tables.h" - -/** - @ingroup groupFastMath - */ - -/** - @addtogroup SQRT - @{ - */ - -/** - @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 ARM_MATH_SUCCESS : input value is positive - - \ref ARM_MATH_ARGUMENT_ERROR : input value is negative; *pOut is set to 0 - */ -#define Q28QUARTER 0x20000000 - -arm_status arm_sqrt_q31( - q31_t in, - q31_t * pOut) -{ - q31_t number, var1, signBits1 ,temp; - - number = in; - - /* If the input is a positive number then compute the signBits. */ - if (number > 0) - { - signBits1 = __CLZ(number) - 1; - - /* Shift by the number of signBits1 */ - if ((signBits1 % 2) == 0) - { - number = number << signBits1; - } - else - { - number = number << (signBits1 - 1); - } - - /* Start value for 1/sqrt(x) for the Newton iteration */ - var1 = sqrt_initial_lut_q31[(number>> 26) - (Q28QUARTER >> 26)]; - - /* 0.5 var1 * (3 - number * var1 * var1) */ - - /* 1st iteration */ - - temp = ((q63_t) var1 * var1) >> 28; - temp = ((q63_t) number * temp) >> 31; - temp = 0x30000000 - temp; - var1 = ((q63_t) var1 * temp) >> 29; - - - /* 2nd iteration */ - temp = ((q63_t) var1 * var1) >> 28; - temp = ((q63_t) number * temp) >> 31; - temp = 0x30000000 - temp; - var1 = ((q63_t) var1 * temp) >> 29; - - /* 3nd iteration */ - temp = ((q63_t) var1 * var1) >> 28; - temp = ((q63_t) number * temp) >> 31; - temp = 0x30000000 - temp; - var1 = ((q63_t) var1 * temp) >> 29; - - /* Multiply the inverse square root with the original value */ - var1 = ((q31_t) (((q63_t) number * var1) >> 28)); - - /* Shift the output down accordingly */ - if ((signBits1 % 2) == 0) - { - var1 = var1 >> (signBits1 / 2); - } - else - { - var1 = var1 >> ((signBits1 - 1) / 2); - } - *pOut = var1; - - return (ARM_MATH_SUCCESS); - } - /* If the number is a negative number then store zero as its square root value */ - else - { - *pOut = 0; - - return (ARM_MATH_ARGUMENT_ERROR); - } -} - -/** - @} end of SQRT group - */ +/* ---------------------------------------------------------------------- + * Project: CMSIS DSP Library + * Title: arm_sqrt_q31.c + * Description: Q31 square root function + * + * $Date: 18. March 2019 + * $Revision: V1.6.0 + * + * Target Processor: Cortex-M cores + * -------------------------------------------------------------------- */ +/* + * Copyright (C) 2010-2019 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_math.h" +#include "arm_common_tables.h" + +/** + @ingroup groupFastMath + */ + +/** + @addtogroup SQRT + @{ + */ + +/** + @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 ARM_MATH_SUCCESS : input value is positive + - \ref ARM_MATH_ARGUMENT_ERROR : input value is negative; *pOut is set to 0 + */ + +arm_status arm_sqrt_q31( + q31_t in, + q31_t * pOut) +{ + q31_t bits_val1; + q31_t number, temp1, var1, signBits1, half; + float32_t temp_float1; + union + { + q31_t fracval; + float32_t floatval; + } tempconv; + + number = in; + + /* If the input is a positive number then compute the signBits. */ + if (number > 0) + { + signBits1 = __CLZ(number) - 1; + + /* Shift by the number of signBits1 */ + if ((signBits1 % 2) == 0) + { + number = number << signBits1; + } + else + { + number = number << (signBits1 - 1); + } + + /* Calculate half value of the number */ + half = number >> 1; + /* Store the number for later use */ + temp1 = number; + + /* Convert to float */ + temp_float1 = number * 4.6566128731e-010f; + /* Store as integer */ + tempconv.floatval = temp_float1; + bits_val1 = tempconv.fracval; + /* Subtract the shifted value from the magic number to give intial guess */ + bits_val1 = 0x5f3759df - (bits_val1 >> 1); /* gives initial guess */ + /* Store as float */ + tempconv.fracval = bits_val1; + temp_float1 = tempconv.floatval; + /* Convert to integer format */ + var1 = (q31_t) (temp_float1 * 1073741824); + + /* 1st iteration */ + var1 = ((q31_t) ((q63_t) var1 * (0x30000000 - + ((q31_t) + ((((q31_t) + (((q63_t) var1 * var1) >> 31)) * + (q63_t) half) >> 31))) >> 31)) << 2; + /* 2nd iteration */ + var1 = ((q31_t) ((q63_t) var1 * (0x30000000 - + ((q31_t) + ((((q31_t) + (((q63_t) var1 * var1) >> 31)) * + (q63_t) half) >> 31))) >> 31)) << 2; + /* 3rd iteration */ + var1 = ((q31_t) ((q63_t) var1 * (0x30000000 - + ((q31_t) + ((((q31_t) + (((q63_t) var1 * var1) >> 31)) * + (q63_t) half) >> 31))) >> 31)) << 2; + + /* Multiply the inverse square root with the original value */ + var1 = ((q31_t) (((q63_t) temp1 * var1) >> 31)) << 1; + + /* Shift the output down accordingly */ + if ((signBits1 % 2) == 0) + { + var1 = var1 >> (signBits1 / 2); + } + else + { + var1 = var1 >> ((signBits1 - 1) / 2); + } + *pOut = var1; + + return (ARM_MATH_SUCCESS); + } + /* If the number is a negative number then store zero as its square root value */ + else + { + *pOut = 0; + + return (ARM_MATH_ARGUMENT_ERROR); + } +} + +/** + @} end of SQRT group + */ -- cgit v1.2.3