tmk_core/tool/mbed/mbed-sdk/libraries/dsp/cmsis_dsp/BasicMathFunctions/arm_dot_prod_q15.c

   1 /* ----------------------------------------------------------------------
   2 * Copyright (C) 2010-2013 ARM Limited. All rights reserved.
   3 *
   4 * $Date:        17. January 2013
   5 * $Revision:    V1.4.1
   6 *
   7 * Project:          CMSIS DSP Library
   8 * Title:                arm_dot_prod_q15.c
   9 *
  10 * Description:  Q15 dot product.
  11 *
  12 * Target Processor: Cortex-M4/Cortex-M3/Cortex-M0
  13 *
  14 * Redistribution and use in source and binary forms, with or without
  15 * modification, are permitted provided that the following conditions
  16 * are met:
  17 *   - Redistributions of source code must retain the above copyright
  18 *     notice, this list of conditions and the following disclaimer.
  19 *   - Redistributions in binary form must reproduce the above copyright
  20 *     notice, this list of conditions and the following disclaimer in
  21 *     the documentation and/or other materials provided with the
  22 *     distribution.
  23 *   - Neither the name of ARM LIMITED nor the names of its contributors
  24 *     may be used to endorse or promote products derived from this
  25 *     software without specific prior written permission.
  26 *
  27 * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
  28 * "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
  29 * LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS
  30 * FOR A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE
  31 * COPYRIGHT OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT,
  32 * INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING,
  33 * BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES;
  34 * LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER
  35 * CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
  36 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN
  37 * ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE
  38 * POSSIBILITY OF SUCH DAMAGE.
  39 * -------------------------------------------------------------------- */
  40
  41 #include "arm_math.h"
  42
  43 /**
  44  * @ingroup groupMath
  45  */
  46
  47 /**
  48  * @addtogroup dot_prod
  49  * @{
  50  */
  51
  52 /**
  53  * @brief Dot product of Q15 vectors.
  54  * @param[in]       *pSrcA points to the first input vector
  55  * @param[in]       *pSrcB points to the second input vector
  56  * @param[in]       blockSize number of samples in each vector
  57  * @param[out]      *result output result returned here
  58  * @return none.
  59  *
  60  * <b>Scaling and Overflow Behavior:</b>
  61  * \par
  62  * The intermediate multiplications are in 1.15 x 1.15 = 2.30 format and these
  63  * results are added to a 64-bit accumulator in 34.30 format.
  64  * Nonsaturating additions are used and given that there are 33 guard bits in the accumulator
  65  * there is no risk of overflow.
  66  * The return result is in 34.30 format.
  67  */
  68
  69 void arm_dot_prod_q15(
  70   q15_t * pSrcA,
  71   q15_t * pSrcB,
  72   uint32_t blockSize,
  73   q63_t * result)
  74 {
  75   q63_t sum = 0;                                 /* Temporary result storage */
  76   uint32_t blkCnt;                               /* loop counter */
  77
  78 #ifndef ARM_MATH_CM0_FAMILY
  79
  80 /* Run the below code for Cortex-M4 and Cortex-M3 */
  81
  82
  83   /*loop Unrolling */
  84   blkCnt = blockSize >> 2u;
  85
  86   /* First part of the processing with loop unrolling.  Compute 4 outputs at a time.
  87    ** a second loop below computes the remaining 1 to 3 samples. */
  88   while(blkCnt > 0u)
  89   {
  90     /* C = A[0]* B[0] + A[1]* B[1] + A[2]* B[2] + .....+ A[blockSize-1]* B[blockSize-1] */
  91     /* Calculate dot product and then store the result in a temporary buffer. */
  92     sum = __SMLALD(*__SIMD32(pSrcA)++, *__SIMD32(pSrcB)++, sum);
  93     sum = __SMLALD(*__SIMD32(pSrcA)++, *__SIMD32(pSrcB)++, sum);
  94
  95     /* Decrement the loop counter */
  96     blkCnt--;
  97   }
  98
  99   /* If the blockSize is not a multiple of 4, compute any remaining output samples here.
 100    ** No loop unrolling is used. */
 101   blkCnt = blockSize % 0x4u;
 102
 103   while(blkCnt > 0u)
 104   {
 105     /* C = A[0]* B[0] + A[1]* B[1] + A[2]* B[2] + .....+ A[blockSize-1]* B[blockSize-1] */
 106     /* Calculate dot product and then store the results in a temporary buffer. */
 107     sum = __SMLALD(*pSrcA++, *pSrcB++, sum);
 108
 109     /* Decrement the loop counter */
 110     blkCnt--;
 111   }
 112
 113
 114 #else
 115
 116   /* Run the below code for Cortex-M0 */
 117
 118   /* Initialize blkCnt with number of samples */
 119   blkCnt = blockSize;
 120
 121   while(blkCnt > 0u)
 122   {
 123     /* C = A[0]* B[0] + A[1]* B[1] + A[2]* B[2] + .....+ A[blockSize-1]* B[blockSize-1] */
 124     /* Calculate dot product and then store the results in a temporary buffer. */
 125     sum += (q63_t) ((q31_t) * pSrcA++ * *pSrcB++);
 126
 127     /* Decrement the loop counter */
 128     blkCnt--;
 129   }
 130
 131 #endif /* #ifndef ARM_MATH_CM0_FAMILY */
 132
 133   /* Store the result in the destination buffer in 34.30 format */
 134   *result = sum;
 135
 136 }
 137
 138 /**
 139  * @} end of dot_prod group
 140  */