lib/main/CMSIS/DSP/Source/FastMathFunctions/arm_sqrt_q31.c

   1 /* ----------------------------------------------------------------------
   2  * Project:      CMSIS DSP Library
   3  * Title:        arm_sqrt_q31.c
   4  * Description:  Q31 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  * @ingroup groupFastMath
  34  */
  35
  36 /**
  37  * @addtogroup SQRT
  38  * @{
  39  */
  40
  41 /**
  42  * @brief Q31 square root function.
  43  * @param[in]   in    input value.  The range of the input value is [0 +1) or 0x00000000 to 0x7FFFFFFF.
  44  * @param[out]  *pOut square root of input value.
  45  * @return The function returns ARM_MATH_SUCCESS if the input value is positive
  46  * and ARM_MATH_ARGUMENT_ERROR if the input is negative.  For
  47  * negative inputs, the function returns *pOut = 0.
  48  */
  49
  50 arm_status arm_sqrt_q31(
  51   q31_t in,
  52   q31_t * pOut)
  53 {
  54   q31_t number, temp1, bits_val1, var1, signBits1, half;
  55   float32_t temp_float1;
  56   union
  57   {
  58       q31_t fracval;
  59       float32_t floatval;
  60   } tempconv;
  61
  62   number = in;
  63
  64   /* If the input is a positive number then compute the signBits. */
  65   if (number > 0)
  66   {
  67     signBits1 = __CLZ(number) - 1;
  68
  69     /* Shift by the number of signBits1 */
  70     if ((signBits1 % 2) == 0)
  71     {
  72       number = number << signBits1;
  73     }
  74     else
  75     {
  76       number = number << (signBits1 - 1);
  77     }
  78
  79     /* Calculate half value of the number */
  80     half = number >> 1;
  81     /* Store the number for later use */
  82     temp1 = number;
  83
  84     /*Convert to float */
  85     temp_float1 = number * 4.6566128731e-010f;
  86     /*Store as integer */
  87     tempconv.floatval = temp_float1;
  88     bits_val1 = tempconv.fracval;
  89     /* Subtract the shifted value from the magic number to give intial guess */
  90     bits_val1 = 0x5f3759df - (bits_val1 >> 1);  /* gives initial guess */
  91     /* Store as float */
  92     tempconv.fracval = bits_val1;
  93     temp_float1 = tempconv.floatval;
  94     /* Convert to integer format */
  95     var1 = (q31_t) (temp_float1 * 1073741824);
  96
  97     /* 1st iteration */
  98     var1 = ((q31_t) ((q63_t) var1 * (0x30000000 -
  99                                      ((q31_t)
 100                                       ((((q31_t)
 101                                          (((q63_t) var1 * var1) >> 31)) *
 102                                         (q63_t) half) >> 31))) >> 31)) << 2;
 103     /* 2nd iteration */
 104     var1 = ((q31_t) ((q63_t) var1 * (0x30000000 -
 105                                      ((q31_t)
 106                                       ((((q31_t)
 107                                          (((q63_t) var1 * var1) >> 31)) *
 108                                         (q63_t) half) >> 31))) >> 31)) << 2;
 109     /* 3rd iteration */
 110     var1 = ((q31_t) ((q63_t) var1 * (0x30000000 -
 111                                      ((q31_t)
 112                                       ((((q31_t)
 113                                          (((q63_t) var1 * var1) >> 31)) *
 114                                         (q63_t) half) >> 31))) >> 31)) << 2;
 115
 116     /* Multiply the inverse square root with the original value */
 117     var1 = ((q31_t) (((q63_t) temp1 * var1) >> 31)) << 1;
 118
 119     /* Shift the output down accordingly */
 120     if ((signBits1 % 2) == 0)
 121     {
 122       var1 = var1 >> (signBits1 / 2);
 123     }
 124     else
 125     {
 126       var1 = var1 >> ((signBits1 - 1) / 2);
 127     }
 128     *pOut = var1;
 129
 130     return (ARM_MATH_SUCCESS);
 131   }
 132   /* If the number is a negative number then store zero as its square root value */
 133   else
 134   {
 135     *pOut = 0;
 136     return (ARM_MATH_ARGUMENT_ERROR);
 137   }
 138 }
 139
 140 /**
 141  * @} end of SQRT group
 142  */