lib/main/CMSIS/DSP/Source/ComplexMathFunctions/arm_cmplx_mag_q31.c

   1 /* ----------------------------------------------------------------------
   2  * Project:      CMSIS DSP Library
   3  * Title:        arm_cmplx_mag_q31.c
   4  * Description:  Q31 complex magnitude
   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 groupCmplxMath
  33  */
  34
  35 /**
  36  * @addtogroup cmplx_mag
  37  * @{
  38  */
  39
  40 /**
  41  * @brief  Q31 complex magnitude
  42  * @param  *pSrc points to the complex input vector
  43  * @param  *pDst points to the real output vector
  44  * @param  numSamples number of complex samples in the input vector
  45  * @return none.
  46  *
  47  * <b>Scaling and Overflow Behavior:</b>
  48  * \par
  49  * The function implements 1.31 by 1.31 multiplications and finally output is converted into 2.30 format.
  50  * Input down scaling is not required.
  51  */
  52
  53 void arm_cmplx_mag_q31(
  54   q31_t * pSrc,
  55   q31_t * pDst,
  56   uint32_t numSamples)
  57 {
  58   q31_t real, imag;                              /* Temporary variables to hold input values */
  59   q31_t acc0, acc1;                              /* Accumulators */
  60   uint32_t blkCnt;                               /* loop counter */
  61
  62 #if defined (ARM_MATH_DSP)
  63
  64   /* Run the below code for Cortex-M4 and Cortex-M3 */
  65   q31_t real1, real2, imag1, imag2;              /* Temporary variables to hold input values */
  66   q31_t out1, out2, out3, out4;                  /* Accumulators */
  67   q63_t mul1, mul2, mul3, mul4;                  /* Temporary variables */
  68
  69
  70   /*loop Unrolling */
  71   blkCnt = numSamples >> 2U;
  72
  73   /* First part of the processing with loop unrolling.  Compute 4 outputs at a time.
  74    ** a second loop below computes the remaining 1 to 3 samples. */
  75   while (blkCnt > 0U)
  76   {
  77     /* read complex input from source buffer */
  78     real1 = pSrc[0];
  79     imag1 = pSrc[1];
  80     real2 = pSrc[2];
  81     imag2 = pSrc[3];
  82
  83     /* calculate power of input values */
  84     mul1 = (q63_t) real1 *real1;
  85     mul2 = (q63_t) imag1 *imag1;
  86     mul3 = (q63_t) real2 *real2;
  87     mul4 = (q63_t) imag2 *imag2;
  88
  89     /* get the result to 3.29 format */
  90     out1 = (q31_t) (mul1 >> 33);
  91     out2 = (q31_t) (mul2 >> 33);
  92     out3 = (q31_t) (mul3 >> 33);
  93     out4 = (q31_t) (mul4 >> 33);
  94
  95     /* add real and imaginary accumulators */
  96     out1 = out1 + out2;
  97     out3 = out3 + out4;
  98
  99     /* read complex input from source buffer */
 100     real1 = pSrc[4];
 101     imag1 = pSrc[5];
 102     real2 = pSrc[6];
 103     imag2 = pSrc[7];
 104
 105     /* calculate square root */
 106     arm_sqrt_q31(out1, &pDst[0]);
 107
 108     /* calculate power of input values */
 109     mul1 = (q63_t) real1 *real1;
 110
 111     /* calculate square root */
 112     arm_sqrt_q31(out3, &pDst[1]);
 113
 114     /* calculate power of input values */
 115     mul2 = (q63_t) imag1 *imag1;
 116     mul3 = (q63_t) real2 *real2;
 117     mul4 = (q63_t) imag2 *imag2;
 118
 119     /* get the result to 3.29 format */
 120     out1 = (q31_t) (mul1 >> 33);
 121     out2 = (q31_t) (mul2 >> 33);
 122     out3 = (q31_t) (mul3 >> 33);
 123     out4 = (q31_t) (mul4 >> 33);
 124
 125     /* add real and imaginary accumulators */
 126     out1 = out1 + out2;
 127     out3 = out3 + out4;
 128
 129     /* calculate square root */
 130     arm_sqrt_q31(out1, &pDst[2]);
 131
 132     /* increment destination by 8 to process next samples */
 133     pSrc += 8U;
 134
 135     /* calculate square root */
 136     arm_sqrt_q31(out3, &pDst[3]);
 137
 138     /* increment destination by 4 to process next samples */
 139     pDst += 4U;
 140
 141     /* Decrement the loop counter */
 142     blkCnt--;
 143   }
 144
 145   /* If the numSamples is not a multiple of 4, compute any remaining output samples here.
 146    ** No loop unrolling is used. */
 147   blkCnt = numSamples % 0x4U;
 148
 149 #else
 150
 151   /* Run the below code for Cortex-M0 */
 152   blkCnt = numSamples;
 153
 154 #endif /* #if defined (ARM_MATH_DSP) */
 155
 156   while (blkCnt > 0U)
 157   {
 158     /* C[0] = sqrt(A[0] * A[0] + A[1] * A[1]) */
 159     real = *pSrc++;
 160     imag = *pSrc++;
 161     acc0 = (q31_t) (((q63_t) real * real) >> 33);
 162     acc1 = (q31_t) (((q63_t) imag * imag) >> 33);
 163     /* store the result in 2.30 format in the destination buffer. */
 164     arm_sqrt_q31(acc0 + acc1, pDst++);
 165
 166     /* Decrement the loop counter */
 167     blkCnt--;
 168   }
 169 }
 170
 171 /**
 172  * @} end of cmplx_mag group
 173  */