lib/main/CMSIS/DSP/Source/ComplexMathFunctions/arm_cmplx_mult_cmplx_q15.c

   1 /* ----------------------------------------------------------------------
   2  * Project:      CMSIS DSP Library
   3  * Title:        arm_cmplx_mult_cmplx_q15.c
   4  * Description:  Q15 complex-by-complex multiplication
   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 CmplxByCmplxMult
  37  * @{
  38  */
  39
  40 /**
  41  * @brief  Q15 complex-by-complex multiplication
  42  * @param[in]  *pSrcA points to the first input vector
  43  * @param[in]  *pSrcB points to the second input vector
  44  * @param[out]  *pDst  points to the output vector
  45  * @param[in]  numSamples number of complex samples in each vector
  46  * @return none.
  47  *
  48  * <b>Scaling and Overflow Behavior:</b>
  49  * \par
  50  * The function implements 1.15 by 1.15 multiplications and finally output is converted into 3.13 format.
  51  */
  52
  53 void arm_cmplx_mult_cmplx_q15(
  54   q15_t * pSrcA,
  55   q15_t * pSrcB,
  56   q15_t * pDst,
  57   uint32_t numSamples)
  58 {
  59   q15_t a, b, c, d;                              /* Temporary variables to store real and imaginary values */
  60
  61 #if defined (ARM_MATH_DSP)
  62
  63   /* Run the below code for Cortex-M4 and Cortex-M3 */
  64   uint32_t blkCnt;                               /* loop counters */
  65
  66   /* loop Unrolling */
  67   blkCnt = numSamples >> 2U;
  68
  69   /* First part of the processing with loop unrolling.  Compute 4 outputs at a time.
  70    ** a second loop below computes the remaining 1 to 3 samples. */
  71   while (blkCnt > 0U)
  72   {
  73     /* C[2 * i] = A[2 * i] * B[2 * i] - A[2 * i + 1] * B[2 * i + 1].  */
  74     /* C[2 * i + 1] = A[2 * i] * B[2 * i + 1] + A[2 * i + 1] * B[2 * i].  */
  75     a = *pSrcA++;
  76     b = *pSrcA++;
  77     c = *pSrcB++;
  78     d = *pSrcB++;
  79
  80     /* store the result in 3.13 format in the destination buffer. */
  81     *pDst++ =
  82       (q15_t) (q31_t) (((q31_t) a * c) >> 17) - (((q31_t) b * d) >> 17);
  83     /* store the result in 3.13 format in the destination buffer. */
  84     *pDst++ =
  85       (q15_t) (q31_t) (((q31_t) a * d) >> 17) + (((q31_t) b * c) >> 17);
  86
  87     a = *pSrcA++;
  88     b = *pSrcA++;
  89     c = *pSrcB++;
  90     d = *pSrcB++;
  91
  92     /* store the result in 3.13 format in the destination buffer. */
  93     *pDst++ =
  94       (q15_t) (q31_t) (((q31_t) a * c) >> 17) - (((q31_t) b * d) >> 17);
  95     /* store the result in 3.13 format in the destination buffer. */
  96     *pDst++ =
  97       (q15_t) (q31_t) (((q31_t) a * d) >> 17) + (((q31_t) b * c) >> 17);
  98
  99     a = *pSrcA++;
 100     b = *pSrcA++;
 101     c = *pSrcB++;
 102     d = *pSrcB++;
 103
 104     /* store the result in 3.13 format in the destination buffer. */
 105     *pDst++ =
 106       (q15_t) (q31_t) (((q31_t) a * c) >> 17) - (((q31_t) b * d) >> 17);
 107     /* store the result in 3.13 format in the destination buffer. */
 108     *pDst++ =
 109       (q15_t) (q31_t) (((q31_t) a * d) >> 17) + (((q31_t) b * c) >> 17);
 110
 111     a = *pSrcA++;
 112     b = *pSrcA++;
 113     c = *pSrcB++;
 114     d = *pSrcB++;
 115
 116     /* store the result in 3.13 format in the destination buffer. */
 117     *pDst++ =
 118       (q15_t) (q31_t) (((q31_t) a * c) >> 17) - (((q31_t) b * d) >> 17);
 119     /* store the result in 3.13 format in the destination buffer. */
 120     *pDst++ =
 121       (q15_t) (q31_t) (((q31_t) a * d) >> 17) + (((q31_t) b * c) >> 17);
 122
 123     /* Decrement the blockSize loop counter */
 124     blkCnt--;
 125   }
 126
 127   /* If the blockSize is not a multiple of 4, compute any remaining output samples here.
 128    ** No loop unrolling is used. */
 129   blkCnt = numSamples % 0x4U;
 130
 131   while (blkCnt > 0U)
 132   {
 133     /* C[2 * i] = A[2 * i] * B[2 * i] - A[2 * i + 1] * B[2 * i + 1].  */
 134     /* C[2 * i + 1] = A[2 * i] * B[2 * i + 1] + A[2 * i + 1] * B[2 * i].  */
 135     a = *pSrcA++;
 136     b = *pSrcA++;
 137     c = *pSrcB++;
 138     d = *pSrcB++;
 139
 140     /* store the result in 3.13 format in the destination buffer. */
 141     *pDst++ =
 142       (q15_t) (q31_t) (((q31_t) a * c) >> 17) - (((q31_t) b * d) >> 17);
 143     /* store the result in 3.13 format in the destination buffer. */
 144     *pDst++ =
 145       (q15_t) (q31_t) (((q31_t) a * d) >> 17) + (((q31_t) b * c) >> 17);
 146
 147     /* Decrement the blockSize loop counter */
 148     blkCnt--;
 149   }
 150
 151 #else
 152
 153   /* Run the below code for Cortex-M0 */
 154
 155   while (numSamples > 0U)
 156   {
 157     /* C[2 * i] = A[2 * i] * B[2 * i] - A[2 * i + 1] * B[2 * i + 1].  */
 158     /* C[2 * i + 1] = A[2 * i] * B[2 * i + 1] + A[2 * i + 1] * B[2 * i].  */
 159     a = *pSrcA++;
 160     b = *pSrcA++;
 161     c = *pSrcB++;
 162     d = *pSrcB++;
 163
 164     /* store the result in 3.13 format in the destination buffer. */
 165     *pDst++ =
 166       (q15_t) (q31_t) (((q31_t) a * c) >> 17) - (((q31_t) b * d) >> 17);
 167     /* store the result in 3.13 format in the destination buffer. */
 168     *pDst++ =
 169       (q15_t) (q31_t) (((q31_t) a * d) >> 17) + (((q31_t) b * c) >> 17);
 170
 171     /* Decrement the blockSize loop counter */
 172     numSamples--;
 173   }
 174
 175 #endif /* #if defined (ARM_MATH_DSP) */
 176
 177 }
 178
 179 /**
 180  * @} end of CmplxByCmplxMult group
 181  */