Details | Last modification | View Log | RSS feed
| Rev | Author | Line No. | Line |
|---|---|---|---|
| 56 | mjames | 1 | /* ---------------------------------------------------------------------- |
| 2 | * Project: CMSIS DSP Library |
||
| 3 | * Title: arm_sqrt_q15.c |
||
| 4 | * Description: Q15 square root function |
||
| 5 | * |
||
| 6 | * $Date: 27. January 2017 |
||
| 7 | * $Revision: V.1.5.1 |
||
| 8 | * |
||
| 9 | * Target Processor: Cortex-M cores |
||
| 10 | * -------------------------------------------------------------------- */ |
||
| 11 | /* |
||
| 12 | * Copyright (C) 2010-2017 ARM Limited or its affiliates. All rights reserved. |
||
| 13 | * |
||
| 14 | * SPDX-License-Identifier: Apache-2.0 |
||
| 15 | * |
||
| 16 | * Licensed under the Apache License, Version 2.0 (the License); you may |
||
| 17 | * not use this file except in compliance with the License. |
||
| 18 | * You may obtain a copy of the License at |
||
| 19 | * |
||
| 20 | * www.apache.org/licenses/LICENSE-2.0 |
||
| 21 | * |
||
| 22 | * Unless required by applicable law or agreed to in writing, software |
||
| 23 | * distributed under the License is distributed on an AS IS BASIS, WITHOUT |
||
| 24 | * WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. |
||
| 25 | * See the License for the specific language governing permissions and |
||
| 26 | * limitations under the License. |
||
| 27 | */ |
||
| 28 | |||
| 29 | #include "arm_math.h" |
||
| 30 | #include "arm_common_tables.h" |
||
| 31 | |||
| 32 | |||
| 33 | /** |
||
| 34 | * @ingroup groupFastMath |
||
| 35 | */ |
||
| 36 | |||
| 37 | /** |
||
| 38 | * @addtogroup SQRT |
||
| 39 | * @{ |
||
| 40 | */ |
||
| 41 | |||
| 42 | /** |
||
| 43 | * @brief Q15 square root function. |
||
| 44 | * @param[in] in input value. The range of the input value is [0 +1) or 0x0000 to 0x7FFF. |
||
| 45 | * @param[out] *pOut square root of input value. |
||
| 46 | * @return The function returns ARM_MATH_SUCCESS if the input value is positive |
||
| 47 | * and ARM_MATH_ARGUMENT_ERROR if the input is negative. For |
||
| 48 | * negative inputs, the function returns *pOut = 0. |
||
| 49 | */ |
||
| 50 | |||
| 51 | arm_status arm_sqrt_q15( |
||
| 52 | q15_t in, |
||
| 53 | q15_t * pOut) |
||
| 54 | { |
||
| 55 | q15_t number, temp1, var1, signBits1, half; |
||
| 56 | q31_t bits_val1; |
||
| 57 | float32_t temp_float1; |
||
| 58 | union |
||
| 59 | { |
||
| 60 | q31_t fracval; |
||
| 61 | float32_t floatval; |
||
| 62 | } tempconv; |
||
| 63 | |||
| 64 | number = in; |
||
| 65 | |||
| 66 | /* If the input is a positive number then compute the signBits. */ |
||
| 67 | if (number > 0) |
||
| 68 | { |
||
| 69 | signBits1 = __CLZ(number) - 17; |
||
| 70 | |||
| 71 | /* Shift by the number of signBits1 */ |
||
| 72 | if ((signBits1 % 2) == 0) |
||
| 73 | { |
||
| 74 | number = number << signBits1; |
||
| 75 | } |
||
| 76 | else |
||
| 77 | { |
||
| 78 | number = number << (signBits1 - 1); |
||
| 79 | } |
||
| 80 | |||
| 81 | /* Calculate half value of the number */ |
||
| 82 | half = number >> 1; |
||
| 83 | /* Store the number for later use */ |
||
| 84 | temp1 = number; |
||
| 85 | |||
| 86 | /* Convert to float */ |
||
| 87 | temp_float1 = number * 3.051757812500000e-005f; |
||
| 88 | /*Store as integer */ |
||
| 89 | tempconv.floatval = temp_float1; |
||
| 90 | bits_val1 = tempconv.fracval; |
||
| 91 | /* Subtract the shifted value from the magic number to give intial guess */ |
||
| 92 | bits_val1 = 0x5f3759df - (bits_val1 >> 1); /* gives initial guess */ |
||
| 93 | /* Store as float */ |
||
| 94 | tempconv.fracval = bits_val1; |
||
| 95 | temp_float1 = tempconv.floatval; |
||
| 96 | /* Convert to integer format */ |
||
| 97 | var1 = (q31_t) (temp_float1 * 16384); |
||
| 98 | |||
| 99 | /* 1st iteration */ |
||
| 100 | var1 = ((q15_t) ((q31_t) var1 * (0x3000 - |
||
| 101 | ((q15_t) |
||
| 102 | ((((q15_t) |
||
| 103 | (((q31_t) var1 * var1) >> 15)) * |
||
| 104 | (q31_t) half) >> 15))) >> 15)) << 2; |
||
| 105 | /* 2nd iteration */ |
||
| 106 | var1 = ((q15_t) ((q31_t) var1 * (0x3000 - |
||
| 107 | ((q15_t) |
||
| 108 | ((((q15_t) |
||
| 109 | (((q31_t) var1 * var1) >> 15)) * |
||
| 110 | (q31_t) half) >> 15))) >> 15)) << 2; |
||
| 111 | /* 3rd iteration */ |
||
| 112 | var1 = ((q15_t) ((q31_t) var1 * (0x3000 - |
||
| 113 | ((q15_t) |
||
| 114 | ((((q15_t) |
||
| 115 | (((q31_t) var1 * var1) >> 15)) * |
||
| 116 | (q31_t) half) >> 15))) >> 15)) << 2; |
||
| 117 | |||
| 118 | /* Multiply the inverse square root with the original value */ |
||
| 119 | var1 = ((q15_t) (((q31_t) temp1 * var1) >> 15)) << 1; |
||
| 120 | |||
| 121 | /* Shift the output down accordingly */ |
||
| 122 | if ((signBits1 % 2) == 0) |
||
| 123 | { |
||
| 124 | var1 = var1 >> (signBits1 / 2); |
||
| 125 | } |
||
| 126 | else |
||
| 127 | { |
||
| 128 | var1 = var1 >> ((signBits1 - 1) / 2); |
||
| 129 | } |
||
| 130 | *pOut = var1; |
||
| 131 | |||
| 132 | return (ARM_MATH_SUCCESS); |
||
| 133 | } |
||
| 134 | /* If the number is a negative number then store zero as its square root value */ |
||
| 135 | else |
||
| 136 | { |
||
| 137 | *pOut = 0; |
||
| 138 | return (ARM_MATH_ARGUMENT_ERROR); |
||
| 139 | } |
||
| 140 | } |
||
| 141 | |||
| 142 | /** |
||
| 143 | * @} end of SQRT group |
||
| 144 | */ |