arm_mat_scale_q31.c 6.6 KB

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960616263646566676869707172737475767778798081828384858687888990919293949596979899100101102103104105106107108109110111112113114115116117118119120121122123124125126127128129130131132133134135136137138139140141142143144145146147148149150151152153154155156157158159160161162163164165166167168169170171172173174175176177178179180181182183184185186187188189190191192193194195196197198199200201202
  1. /* ----------------------------------------------------------------------
  2. * Copyright (C) 2010-2014 ARM Limited. All rights reserved.
  3. *
  4. * $Date: 31. July 2014
  5. * $Revision: V1.4.4
  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. #include "arm_math.h"
  40. /**
  41. * @ingroup groupMatrix
  42. */
  43. /**
  44. * @addtogroup MatrixScale
  45. * @{
  46. */
  47. /**
  48. * @brief Q31 matrix scaling.
  49. * @param[in] *pSrc points to input matrix
  50. * @param[in] scaleFract fractional portion of the scale factor
  51. * @param[in] shift number of bits to shift the result by
  52. * @param[out] *pDst points to output matrix structure
  53. * @return The function returns either
  54. * <code>ARM_MATH_SIZE_MISMATCH</code> or <code>ARM_MATH_SUCCESS</code> based on the outcome of size checking.
  55. *
  56. * @details
  57. * <b>Scaling and Overflow Behavior:</b>
  58. * \par
  59. * The input data <code>*pSrc</code> and <code>scaleFract</code> are in 1.31 format.
  60. * These are multiplied to yield a 2.62 intermediate result and this is shifted with saturation to 1.31 format.
  61. */
  62. arm_status arm_mat_scale_q31(
  63. const arm_matrix_instance_q31 * pSrc,
  64. q31_t scaleFract,
  65. int32_t shift,
  66. arm_matrix_instance_q31 * pDst)
  67. {
  68. q31_t *pIn = pSrc->pData; /* input data matrix pointer */
  69. q31_t *pOut = pDst->pData; /* output data matrix pointer */
  70. uint32_t numSamples; /* total number of elements in the matrix */
  71. int32_t totShift = shift + 1; /* shift to apply after scaling */
  72. uint32_t blkCnt; /* loop counters */
  73. arm_status status; /* status of matrix scaling */
  74. q31_t in1, in2, out1; /* temporary variabels */
  75. #ifndef ARM_MATH_CM0_FAMILY
  76. q31_t in3, in4, out2, out3, out4; /* temporary variables */
  77. #endif // #ifndef ARM_MAT_CM0
  78. #ifdef ARM_MATH_MATRIX_CHECK
  79. /* Check for matrix mismatch */
  80. if((pSrc->numRows != pDst->numRows) || (pSrc->numCols != pDst->numCols))
  81. {
  82. /* Set status as ARM_MATH_SIZE_MISMATCH */
  83. status = ARM_MATH_SIZE_MISMATCH;
  84. }
  85. else
  86. #endif // #ifdef ARM_MATH_MATRIX_CHECK
  87. {
  88. /* Total number of samples in the input matrix */
  89. numSamples = (uint32_t) pSrc->numRows * pSrc->numCols;
  90. #ifndef ARM_MATH_CM0_FAMILY
  91. /* Run the below code for Cortex-M4 and Cortex-M3 */
  92. /* Loop Unrolling */
  93. blkCnt = numSamples >> 2u;
  94. /* First part of the processing with loop unrolling. Compute 4 outputs at a time.
  95. ** a second loop below computes the remaining 1 to 3 samples. */
  96. while(blkCnt > 0u)
  97. {
  98. /* C(m,n) = A(m,n) * k */
  99. /* Read values from input */
  100. in1 = *pIn;
  101. in2 = *(pIn + 1);
  102. in3 = *(pIn + 2);
  103. in4 = *(pIn + 3);
  104. /* multiply input with scaler value */
  105. in1 = ((q63_t) in1 * scaleFract) >> 32;
  106. in2 = ((q63_t) in2 * scaleFract) >> 32;
  107. in3 = ((q63_t) in3 * scaleFract) >> 32;
  108. in4 = ((q63_t) in4 * scaleFract) >> 32;
  109. /* apply shifting */
  110. out1 = in1 << totShift;
  111. out2 = in2 << totShift;
  112. /* saturate the results. */
  113. if(in1 != (out1 >> totShift))
  114. out1 = 0x7FFFFFFF ^ (in1 >> 31);
  115. if(in2 != (out2 >> totShift))
  116. out2 = 0x7FFFFFFF ^ (in2 >> 31);
  117. out3 = in3 << totShift;
  118. out4 = in4 << totShift;
  119. *pOut = out1;
  120. *(pOut + 1) = out2;
  121. if(in3 != (out3 >> totShift))
  122. out3 = 0x7FFFFFFF ^ (in3 >> 31);
  123. if(in4 != (out4 >> totShift))
  124. out4 = 0x7FFFFFFF ^ (in4 >> 31);
  125. *(pOut + 2) = out3;
  126. *(pOut + 3) = out4;
  127. /* update pointers to process next sampels */
  128. pIn += 4u;
  129. pOut += 4u;
  130. /* Decrement the numSamples loop counter */
  131. blkCnt--;
  132. }
  133. /* If the numSamples is not a multiple of 4, compute any remaining output samples here.
  134. ** No loop unrolling is used. */
  135. blkCnt = numSamples % 0x4u;
  136. #else
  137. /* Run the below code for Cortex-M0 */
  138. /* Initialize blkCnt with number of samples */
  139. blkCnt = numSamples;
  140. #endif /* #ifndef ARM_MATH_CM0_FAMILY */
  141. while(blkCnt > 0u)
  142. {
  143. /* C(m,n) = A(m,n) * k */
  144. /* Scale, saturate and then store the results in the destination buffer. */
  145. in1 = *pIn++;
  146. in2 = ((q63_t) in1 * scaleFract) >> 32;
  147. out1 = in2 << totShift;
  148. if(in2 != (out1 >> totShift))
  149. out1 = 0x7FFFFFFF ^ (in2 >> 31);
  150. *pOut++ = out1;
  151. /* Decrement the numSamples loop counter */
  152. blkCnt--;
  153. }
  154. /* Set status as ARM_MATH_SUCCESS */
  155. status = ARM_MATH_SUCCESS;
  156. }
  157. /* Return to application */
  158. return (status);
  159. }
  160. /**
  161. * @} end of MatrixScale group
  162. */