lib/main/CMSIS/DSP/Source/StatisticsFunctions/arm_power_q31.c

   1 /* ----------------------------------------------------------------------
   2  * Project:      CMSIS DSP Library
   3  * Title:        arm_power_q31.c
   4  * Description:  Sum of the squares of the elements of a Q31 vector
   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
  31 /**
  32  * @ingroup groupStats
  33  */
  34
  35 /**
  36  * @addtogroup power
  37  * @{
  38  */
  39
  40 /**
  41  * @brief Sum of the squares of the elements of a Q31 vector.
  42  * @param[in]       *pSrc points to the input vector
  43  * @param[in]       blockSize length of the input vector
  44  * @param[out]      *pResult sum of the squares value returned here
  45  * @return none.
  46  *
  47  * @details
  48  * <b>Scaling and Overflow Behavior:</b>
  49  *
  50  * \par
  51  * The function is implemented using a 64-bit internal accumulator.
  52  * The input is represented in 1.31 format.
  53  * Intermediate multiplication yields a 2.62 format, and this
  54  * result is truncated to 2.48 format by discarding the lower 14 bits.
  55  * The 2.48 result is then added without saturation to a 64-bit accumulator in 16.48 format.
  56  * With 15 guard bits in the accumulator, there is no risk of overflow, and the
  57  * full precision of the intermediate multiplication is preserved.
  58  * Finally, the return result is in 16.48 format.
  59  *
  60  */
  61
  62 void arm_power_q31(
  63   q31_t * pSrc,
  64   uint32_t blockSize,
  65   q63_t * pResult)
  66 {
  67   q63_t sum = 0;                                 /* Temporary result storage */
  68   q31_t in;
  69   uint32_t blkCnt;                               /* loop counter */
  70
  71
  72 #if defined (ARM_MATH_DSP)
  73   /* Run the below code for Cortex-M4 and Cortex-M3 */
  74
  75   /*loop Unrolling */
  76   blkCnt = blockSize >> 2U;
  77
  78   /* First part of the processing with loop unrolling.  Compute 4 outputs at a time.
  79    ** a second loop below computes the remaining 1 to 3 samples. */
  80   while (blkCnt > 0U)
  81   {
  82     /* C = A[0] * A[0] + A[1] * A[1] + A[2] * A[2] + ... + A[blockSize-1] * A[blockSize-1] */
  83     /* Compute Power then shift intermediate results by 14 bits to maintain 16.48 format and then store the result in a temporary variable sum, providing 15 guard bits. */
  84     in = *pSrc++;
  85     sum += ((q63_t) in * in) >> 14U;
  86
  87     in = *pSrc++;
  88     sum += ((q63_t) in * in) >> 14U;
  89
  90     in = *pSrc++;
  91     sum += ((q63_t) in * in) >> 14U;
  92
  93     in = *pSrc++;
  94     sum += ((q63_t) in * in) >> 14U;
  95
  96     /* Decrement the loop counter */
  97     blkCnt--;
  98   }
  99
 100   /* If the blockSize is not a multiple of 4, compute any remaining output samples here.
 101    ** No loop unrolling is used. */
 102   blkCnt = blockSize % 0x4U;
 103
 104 #else
 105   /* Run the below code for Cortex-M0 */
 106
 107   /* Loop over blockSize number of values */
 108   blkCnt = blockSize;
 109
 110 #endif /* #if defined (ARM_MATH_DSP) */
 111
 112   while (blkCnt > 0U)
 113   {
 114     /* C = A[0] * A[0] + A[1] * A[1] + A[2] * A[2] + ... + A[blockSize-1] * A[blockSize-1] */
 115     /* Compute Power and then store the result in a temporary variable, sum. */
 116     in = *pSrc++;
 117     sum += ((q63_t) in * in) >> 14U;
 118
 119     /* Decrement the loop counter */
 120     blkCnt--;
 121   }
 122
 123   /* Store the results in 16.48 format  */
 124   *pResult = sum;
 125 }
 126
 127 /**
 128  * @} end of power group
 129  */