tmk_core/tool/mbed/mbed-sdk/libraries/dsp/cmsis_dsp/MatrixFunctions/arm_mat_scale_q31.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_mat_scale_q31.c
   9 *
  10 * Description:  Multiplies a Q31 matrix by a scalar.
  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 #include "arm_math.h"
  41
  42 /**
  43  * @ingroup groupMatrix
  44  */
  45
  46 /**
  47  * @addtogroup MatrixScale
  48  * @{
  49  */
  50
  51 /**
  52  * @brief Q31 matrix scaling.
  53  * @param[in]       *pSrc points to input matrix
  54  * @param[in]       scaleFract fractional portion of the scale factor
  55  * @param[in]       shift number of bits to shift the result by
  56  * @param[out]      *pDst points to output matrix structure
  57  * @return              The function returns either
  58  * <code>ARM_MATH_SIZE_MISMATCH</code> or <code>ARM_MATH_SUCCESS</code> based on the outcome of size checking.
  59  *
  60  * @details
  61  * <b>Scaling and Overflow Behavior:</b>
  62  * \par
  63  * The input data <code>*pSrc</code> and <code>scaleFract</code> are in 1.31 format.
  64  * These are multiplied to yield a 2.62 intermediate result and this is shifted with saturation to 1.31 format.
  65  */
  66
  67 arm_status arm_mat_scale_q31(
  68   const arm_matrix_instance_q31 * pSrc,
  69   q31_t scaleFract,
  70   int32_t shift,
  71   arm_matrix_instance_q31 * pDst)
  72 {
  73   q31_t *pIn = pSrc->pData;                      /* input data matrix pointer */
  74   q31_t *pOut = pDst->pData;                     /* output data matrix pointer */
  75   uint32_t numSamples;                           /* total number of elements in the matrix */
  76   int32_t totShift = shift + 1;                  /* shift to apply after scaling */
  77   uint32_t blkCnt;                               /* loop counters  */
  78   arm_status status;                             /* status of matrix scaling      */
  79   q31_t in1, in2, out1;                          /* temporary variabels */
  80
  81 #ifndef ARM_MATH_CM0_FAMILY
  82
  83   q31_t in3, in4, out2, out3, out4;              /* temporary variables */
  84
  85 #endif //      #ifndef ARM_MAT_CM0
  86
  87 #ifdef ARM_MATH_MATRIX_CHECK
  88   /* Check for matrix mismatch  */
  89   if((pSrc->numRows != pDst->numRows) || (pSrc->numCols != pDst->numCols))
  90   {
  91     /* Set status as ARM_MATH_SIZE_MISMATCH */
  92     status = ARM_MATH_SIZE_MISMATCH;
  93   }
  94   else
  95 #endif //    #ifdef ARM_MATH_MATRIX_CHECK
  96   {
  97     /* Total number of samples in the input matrix */
  98     numSamples = (uint32_t) pSrc->numRows * pSrc->numCols;
  99
 100 #ifndef ARM_MATH_CM0_FAMILY
 101
 102     /* Run the below code for Cortex-M4 and Cortex-M3 */
 103
 104     /* Loop Unrolling */
 105     blkCnt = numSamples >> 2u;
 106
 107     /* First part of the processing with loop unrolling.  Compute 4 outputs at a time.
 108      ** a second loop below computes the remaining 1 to 3 samples. */
 109     while(blkCnt > 0u)
 110     {
 111       /* C(m,n) = A(m,n) * k */
 112       /* Read values from input */
 113       in1 = *pIn;
 114       in2 = *(pIn + 1);
 115       in3 = *(pIn + 2);
 116       in4 = *(pIn + 3);
 117
 118       /* multiply input with scaler value */
 119       in1 = ((q63_t) in1 * scaleFract) >> 32;
 120       in2 = ((q63_t) in2 * scaleFract) >> 32;
 121       in3 = ((q63_t) in3 * scaleFract) >> 32;
 122       in4 = ((q63_t) in4 * scaleFract) >> 32;
 123
 124       /* apply shifting */
 125       out1 = in1 << totShift;
 126       out2 = in2 << totShift;
 127
 128       /* saturate the results. */
 129       if(in1 != (out1 >> totShift))
 130         out1 = 0x7FFFFFFF ^ (in1 >> 31);
 131
 132       if(in2 != (out2 >> totShift))
 133         out2 = 0x7FFFFFFF ^ (in2 >> 31);
 134
 135       out3 = in3 << totShift;
 136       out4 = in4 << totShift;
 137
 138       *pOut = out1;
 139       *(pOut + 1) = out2;
 140
 141       if(in3 != (out3 >> totShift))
 142         out3 = 0x7FFFFFFF ^ (in3 >> 31);
 143
 144       if(in4 != (out4 >> totShift))
 145         out4 = 0x7FFFFFFF ^ (in4 >> 31);
 146
 147
 148       *(pOut + 2) = out3;
 149       *(pOut + 3) = out4;
 150
 151       /* update pointers to process next sampels */
 152       pIn += 4u;
 153       pOut += 4u;
 154
 155
 156       /* Decrement the numSamples loop counter */
 157       blkCnt--;
 158     }
 159
 160     /* If the numSamples is not a multiple of 4, compute any remaining output samples here.
 161      ** No loop unrolling is used. */
 162     blkCnt = numSamples % 0x4u;
 163
 164 #else
 165
 166     /* Run the below code for Cortex-M0 */
 167
 168     /* Initialize blkCnt with number of samples */
 169     blkCnt = numSamples;
 170
 171 #endif /* #ifndef ARM_MATH_CM0_FAMILY */
 172
 173     while(blkCnt > 0u)
 174     {
 175       /* C(m,n) = A(m,n) * k */
 176       /* Scale, saturate and then store the results in the destination buffer. */
 177       in1 = *pIn++;
 178
 179       in2 = ((q63_t) in1 * scaleFract) >> 32;
 180
 181       out1 = in2 << totShift;
 182
 183       if(in2 != (out1 >> totShift))
 184         out1 = 0x7FFFFFFF ^ (in2 >> 31);
 185
 186       *pOut++ = out1;
 187
 188       /* Decrement the numSamples loop counter */
 189       blkCnt--;
 190     }
 191
 192     /* Set status as ARM_MATH_SUCCESS */
 193     status = ARM_MATH_SUCCESS;
 194   }
 195
 196   /* Return to application */
 197   return (status);
 198 }
 199
 200 /**
 201  * @} end of MatrixScale group
 202  */