lib/main/CMSIS/DSP/Source/BasicMathFunctions/arm_mult_q31.c

   1 /* ----------------------------------------------------------------------
   2  * Project:      CMSIS DSP Library
   3  * Title:        arm_mult_q31.c
   4  * Description:  Q31 vector 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 groupMath
  33  */
  34
  35 /**
  36  * @addtogroup BasicMult
  37  * @{
  38  */
  39
  40 /**
  41  * @brief Q31 vector 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]       blockSize number of samples in each vector
  46  * @return none.
  47  *
  48  * <b>Scaling and Overflow Behavior:</b>
  49  * \par
  50  * The function uses saturating arithmetic.
  51  * Results outside of the allowable Q31 range[0x80000000 0x7FFFFFFF] will be saturated.
  52  */
  53
  54 void arm_mult_q31(
  55   q31_t * pSrcA,
  56   q31_t * pSrcB,
  57   q31_t * pDst,
  58   uint32_t blockSize)
  59 {
  60   uint32_t blkCnt;                               /* loop counters */
  61
  62 #if defined (ARM_MATH_DSP)
  63
  64 /* Run the below code for Cortex-M4 and Cortex-M3 */
  65   q31_t inA1, inA2, inA3, inA4;                  /* temporary input variables */
  66   q31_t inB1, inB2, inB3, inB4;                  /* temporary input variables */
  67   q31_t out1, out2, out3, out4;                  /* temporary output variables */
  68
  69   /* loop Unrolling */
  70   blkCnt = blockSize >> 2U;
  71
  72   /* First part of the processing with loop unrolling.  Compute 4 outputs at a time.
  73    ** a second loop below computes the remaining 1 to 3 samples. */
  74   while (blkCnt > 0U)
  75   {
  76     /* C = A * B */
  77     /* Multiply the inputs and then store the results in the destination buffer. */
  78     inA1 = *pSrcA++;
  79     inA2 = *pSrcA++;
  80     inA3 = *pSrcA++;
  81     inA4 = *pSrcA++;
  82     inB1 = *pSrcB++;
  83     inB2 = *pSrcB++;
  84     inB3 = *pSrcB++;
  85     inB4 = *pSrcB++;
  86
  87     out1 = ((q63_t) inA1 * inB1) >> 32;
  88     out2 = ((q63_t) inA2 * inB2) >> 32;
  89     out3 = ((q63_t) inA3 * inB3) >> 32;
  90     out4 = ((q63_t) inA4 * inB4) >> 32;
  91
  92     out1 = __SSAT(out1, 31);
  93     out2 = __SSAT(out2, 31);
  94     out3 = __SSAT(out3, 31);
  95     out4 = __SSAT(out4, 31);
  96
  97     *pDst++ = out1 << 1U;
  98     *pDst++ = out2 << 1U;
  99     *pDst++ = out3 << 1U;
 100     *pDst++ = out4 << 1U;
 101
 102     /* Decrement the blockSize loop counter */
 103     blkCnt--;
 104   }
 105
 106   /* If the blockSize is not a multiple of 4, compute any remaining output samples here.
 107    ** No loop unrolling is used. */
 108   blkCnt = blockSize % 0x4U;
 109
 110   while (blkCnt > 0U)
 111   {
 112     /* C = A * B */
 113     /* Multiply the inputs and then store the results in the destination buffer. */
 114     inA1 = *pSrcA++;
 115     inB1 = *pSrcB++;
 116     out1 = ((q63_t) inA1 * inB1) >> 32;
 117     out1 = __SSAT(out1, 31);
 118     *pDst++ = out1 << 1U;
 119
 120     /* Decrement the blockSize loop counter */
 121     blkCnt--;
 122   }
 123
 124 #else
 125
 126   /* Run the below code for Cortex-M0 */
 127
 128   /* Initialize blkCnt with number of samples */
 129   blkCnt = blockSize;
 130
 131
 132   while (blkCnt > 0U)
 133   {
 134     /* C = A * B */
 135     /* Multiply the inputs and then store the results in the destination buffer. */
 136     *pDst++ =
 137       (q31_t) clip_q63_to_q31(((q63_t) (*pSrcA++) * (*pSrcB++)) >> 31);
 138
 139     /* Decrement the blockSize loop counter */
 140     blkCnt--;
 141   }
 142
 143 #endif /* #if defined (ARM_MATH_DSP) */
 144 }
 145
 146 /**
 147  * @} end of BasicMult group
 148  */