diff options
Diffstat (limited to 'DSP/Source/FastMathFunctions')
| -rw-r--r-- | DSP/Source/FastMathFunctions/arm_cos_f32.c | 115 | ||||
| -rw-r--r-- | DSP/Source/FastMathFunctions/arm_cos_q15.c | 84 | ||||
| -rw-r--r-- | DSP/Source/FastMathFunctions/arm_cos_q31.c | 84 | ||||
| -rw-r--r-- | DSP/Source/FastMathFunctions/arm_sin_f32.c | 123 | ||||
| -rw-r--r-- | DSP/Source/FastMathFunctions/arm_sin_q15.c | 76 | ||||
| -rw-r--r-- | DSP/Source/FastMathFunctions/arm_sin_q31.c | 75 | ||||
| -rw-r--r-- | DSP/Source/FastMathFunctions/arm_sqrt_q15.c | 144 | ||||
| -rw-r--r-- | DSP/Source/FastMathFunctions/arm_sqrt_q31.c | 142 |
8 files changed, 843 insertions, 0 deletions
diff --git a/DSP/Source/FastMathFunctions/arm_cos_f32.c b/DSP/Source/FastMathFunctions/arm_cos_f32.c new file mode 100644 index 0000000..e604b3c --- /dev/null +++ b/DSP/Source/FastMathFunctions/arm_cos_f32.c @@ -0,0 +1,115 @@ +/* ---------------------------------------------------------------------- + * Project: CMSIS DSP Library + * Title: arm_cos_f32.c + * Description: Fast cosine calculation for floating-point values + * + * $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.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 and in the range [0 2*pi) 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 <code>(1.0f-fract)*a + fract*b;</code> + * + * where + * <pre> + * b=Table[index+0]; + * c=Table[index+1]; + * </pre> + */ + + /** + * @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 variables for input, output */ + 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 the 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) & 0x1ff; + + /* 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 the output value */ + return (cosVal); +} + +/** + * @} end of cos group + */ diff --git a/DSP/Source/FastMathFunctions/arm_cos_q15.c b/DSP/Source/FastMathFunctions/arm_cos_q15.c new file mode 100644 index 0000000..7fa2e18 --- /dev/null +++ b/DSP/Source/FastMathFunctions/arm_cos_q15.c @@ -0,0 +1,84 @@ +/* ---------------------------------------------------------------------- + * Project: CMSIS DSP Library + * Title: arm_cos_q15.c + * Description: Fast cosine calculation for Q15 values + * + * $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.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 variables for input, output */ + int32_t index; /* Index variables */ + q15_t a, b; /* Four 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 cosVal << 1; +} + +/** + * @} end of cos group + */ diff --git a/DSP/Source/FastMathFunctions/arm_cos_q31.c b/DSP/Source/FastMathFunctions/arm_cos_q31.c new file mode 100644 index 0000000..fde5368 --- /dev/null +++ b/DSP/Source/FastMathFunctions/arm_cos_q31.c @@ -0,0 +1,84 @@ +/* ---------------------------------------------------------------------- + * Project: CMSIS DSP Library + * Title: arm_cos_q31.c + * Description: Fast cosine calculation for Q31 values + * + * $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.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 variables for input, output */ + int32_t index; /* Index variables */ + q31_t a, b; /* Four 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 cosVal << 1; +} + +/** + * @} end of cos group + */ diff --git a/DSP/Source/FastMathFunctions/arm_sin_f32.c b/DSP/Source/FastMathFunctions/arm_sin_f32.c new file mode 100644 index 0000000..ce8b9b9 --- /dev/null +++ b/DSP/Source/FastMathFunctions/arm_sin_f32.c @@ -0,0 +1,123 @@ +/* ---------------------------------------------------------------------- + * Project: CMSIS DSP Library + * Title: arm_sin_f32.c + * Description: Fast sine calculation for floating-point values + * + * $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.h" +#include "arm_common_tables.h" +#include <math.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 and in the range [0 2*pi) 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 <code>(1.0f-fract)*a + fract*b;</code> + * + * where + * <pre> + * b=Table[index+0]; + * c=Table[index+1]; + * </pre> + */ + +/** + * @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 variables for input, output */ + uint16_t index; /* Index variable */ + float32_t a, b; /* Two nearest output values */ + int32_t n; + float32_t findex; + + /* Special case for small negative inputs */ + if ((x < 0.0f) && (x >= -1.9e-7f)) { + return x; + } + + /* input x is in radians */ + /* Scale the 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 (x < 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) & 0x1ff; + + /* 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 the output value */ + return (sinVal); +} + +/** + * @} end of sin group + */ diff --git a/DSP/Source/FastMathFunctions/arm_sin_q15.c b/DSP/Source/FastMathFunctions/arm_sin_q15.c new file mode 100644 index 0000000..7c8f627 --- /dev/null +++ b/DSP/Source/FastMathFunctions/arm_sin_q15.c @@ -0,0 +1,76 @@ +/* ---------------------------------------------------------------------- + * Project: CMSIS DSP Library + * Title: arm_sin_q15.c + * Description: Fast sine calculation for Q15 values + * + * $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.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 variables for input, output */ + int32_t index; /* Index variables */ + q15_t a, b; /* Four 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 sinVal << 1; +} + +/** + * @} end of sin group + */ diff --git a/DSP/Source/FastMathFunctions/arm_sin_q31.c b/DSP/Source/FastMathFunctions/arm_sin_q31.c new file mode 100644 index 0000000..8d3c7ac --- /dev/null +++ b/DSP/Source/FastMathFunctions/arm_sin_q31.c @@ -0,0 +1,75 @@ +/* ---------------------------------------------------------------------- + * Project: CMSIS DSP Library + * Title: arm_sin_q31.c + * Description: Fast sine calculation for Q31 values + * + * $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.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 variables */ + q31_t a, b; /* Four 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 sinVal << 1; +} + +/** + * @} end of sin group + */ diff --git a/DSP/Source/FastMathFunctions/arm_sqrt_q15.c b/DSP/Source/FastMathFunctions/arm_sqrt_q15.c new file mode 100644 index 0000000..8487ed3 --- /dev/null +++ b/DSP/Source/FastMathFunctions/arm_sqrt_q15.c @@ -0,0 +1,144 @@ +/* ---------------------------------------------------------------------- + * Project: CMSIS DSP Library + * Title: arm_sqrt_q15.c + * Description: Q15 square root function + * + * $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.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 square root of input value. + * @return The function returns ARM_MATH_SUCCESS if the input value is positive + * and ARM_MATH_ARGUMENT_ERROR if the input is negative. For + * negative inputs, the function returns *pOut = 0. + */ + +arm_status arm_sqrt_q15( + q15_t in, + q15_t * pOut) +{ + q15_t number, temp1, var1, signBits1, half; + q31_t bits_val1; + 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/DSP/Source/FastMathFunctions/arm_sqrt_q31.c b/DSP/Source/FastMathFunctions/arm_sqrt_q31.c new file mode 100644 index 0000000..0deea04 --- /dev/null +++ b/DSP/Source/FastMathFunctions/arm_sqrt_q31.c @@ -0,0 +1,142 @@ +/* ---------------------------------------------------------------------- + * Project: CMSIS DSP Library + * Title: arm_sqrt_q31.c + * Description: Q31 square root function + * + * $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.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 square root of input value. + * @return The function returns ARM_MATH_SUCCESS if the input value is positive + * and ARM_MATH_ARGUMENT_ERROR if the input is negative. For + * negative inputs, the function returns *pOut = 0. + */ + +arm_status arm_sqrt_q31( + q31_t in, + q31_t * pOut) +{ + q31_t number, temp1, bits_val1, 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 + */ |