arm_scale_q15.c 4.3 KB

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960616263646566676869707172737475767778798081828384858687888990919293949596979899100101102103104105106107108109110111112113114115116117118119120121122123124125126127128129130131132133134135136137138139140141142143144
  1. /* ----------------------------------------------------------------------
  2. * Project: CMSIS DSP Library
  3. * Title: arm_scale_q15.c
  4. * Description: Multiplies a Q15 vector by a scalar
  5. *
  6. * $Date: 18. March 2019
  7. * $Revision: V1.6.0
  8. *
  9. * Target Processor: Cortex-M cores
  10. * -------------------------------------------------------------------- */
  11. /*
  12. * Copyright (C) 2010-2019 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. #include "arm_math.h"
  29. /**
  30. @ingroup groupMath
  31. */
  32. /**
  33. @addtogroup BasicScale
  34. @{
  35. */
  36. /**
  37. @brief Multiplies a Q15 vector by a scalar.
  38. @param[in] pSrc points to the input vector
  39. @param[in] scaleFract fractional portion of the scale value
  40. @param[in] shift number of bits to shift the result by
  41. @param[out] pDst points to the output vector
  42. @param[in] blockSize number of samples in each vector
  43. @return none
  44. @par Scaling and Overflow Behavior
  45. The input data <code>*pSrc</code> and <code>scaleFract</code> are in 1.15 format.
  46. These are multiplied to yield a 2.30 intermediate result and this is shifted with saturation to 1.15 format.
  47. */
  48. void arm_scale_q15(
  49. const q15_t *pSrc,
  50. q15_t scaleFract,
  51. int8_t shift,
  52. q15_t *pDst,
  53. uint32_t blockSize)
  54. {
  55. uint32_t blkCnt; /* Loop counter */
  56. int8_t kShift = 15 - shift; /* Shift to apply after scaling */
  57. #if defined (ARM_MATH_LOOPUNROLL)
  58. #if defined (ARM_MATH_DSP)
  59. q31_t inA1, inA2;
  60. q31_t out1, out2, out3, out4; /* Temporary output variables */
  61. q15_t in1, in2, in3, in4; /* Temporary input variables */
  62. #endif
  63. #endif
  64. #if defined (ARM_MATH_LOOPUNROLL)
  65. /* Loop unrolling: Compute 4 outputs at a time */
  66. blkCnt = blockSize >> 2U;
  67. while (blkCnt > 0U)
  68. {
  69. /* C = A * scale */
  70. #if defined (ARM_MATH_DSP)
  71. /* read 2 times 2 samples at a time from source */
  72. inA1 = read_q15x2_ia ((q15_t **) &pSrc);
  73. inA2 = read_q15x2_ia ((q15_t **) &pSrc);
  74. /* Scale inputs and store result in temporary variables
  75. * in single cycle by packing the outputs */
  76. out1 = (q31_t) ((q15_t) (inA1 >> 16) * scaleFract);
  77. out2 = (q31_t) ((q15_t) (inA1 ) * scaleFract);
  78. out3 = (q31_t) ((q15_t) (inA2 >> 16) * scaleFract);
  79. out4 = (q31_t) ((q15_t) (inA2 ) * scaleFract);
  80. /* apply shifting */
  81. out1 = out1 >> kShift;
  82. out2 = out2 >> kShift;
  83. out3 = out3 >> kShift;
  84. out4 = out4 >> kShift;
  85. /* saturate the output */
  86. in1 = (q15_t) (__SSAT(out1, 16));
  87. in2 = (q15_t) (__SSAT(out2, 16));
  88. in3 = (q15_t) (__SSAT(out3, 16));
  89. in4 = (q15_t) (__SSAT(out4, 16));
  90. /* store result to destination */
  91. write_q15x2_ia (&pDst, __PKHBT(in2, in1, 16));
  92. write_q15x2_ia (&pDst, __PKHBT(in4, in3, 16));
  93. #else
  94. *pDst++ = (q15_t) (__SSAT(((q31_t) *pSrc++ * scaleFract) >> kShift, 16));
  95. *pDst++ = (q15_t) (__SSAT(((q31_t) *pSrc++ * scaleFract) >> kShift, 16));
  96. *pDst++ = (q15_t) (__SSAT(((q31_t) *pSrc++ * scaleFract) >> kShift, 16));
  97. *pDst++ = (q15_t) (__SSAT(((q31_t) *pSrc++ * scaleFract) >> kShift, 16));
  98. #endif
  99. /* Decrement loop counter */
  100. blkCnt--;
  101. }
  102. /* Loop unrolling: Compute remaining outputs */
  103. blkCnt = blockSize % 0x4U;
  104. #else
  105. /* Initialize blkCnt with number of samples */
  106. blkCnt = blockSize;
  107. #endif /* #if defined (ARM_MATH_LOOPUNROLL) */
  108. while (blkCnt > 0U)
  109. {
  110. /* C = A * scale */
  111. /* Scale input and store result in destination buffer. */
  112. *pDst++ = (q15_t) (__SSAT(((q31_t) *pSrc++ * scaleFract) >> kShift, 16));
  113. /* Decrement loop counter */
  114. blkCnt--;
  115. }
  116. }
  117. /**
  118. @} end of BasicScale group
  119. */