arm_abs_q15.c 4.3 KB

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960616263646566676869707172737475767778798081828384858687888990919293949596979899100101102103104105106107108109110111112113114115116117118119120121122123124125126127128129130131132133134135136137138139140141142143144145146147148149150151152153154155156157158159160161162163164165166167
  1. /* ----------------------------------------------------------------------
  2. * Project: CMSIS DSP Library
  3. * Title: arm_abs_q15.c
  4. * Description: Q15 vector absolute value
  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. #include "arm_math.h"
  29. /**
  30. * @ingroup groupMath
  31. */
  32. /**
  33. * @addtogroup BasicAbs
  34. * @{
  35. */
  36. /**
  37. * @brief Q15 vector absolute value.
  38. * @param[in] *pSrc points to the input buffer
  39. * @param[out] *pDst points to the output buffer
  40. * @param[in] blockSize number of samples in each vector
  41. * @return none.
  42. *
  43. * <b>Scaling and Overflow Behavior:</b>
  44. * \par
  45. * The function uses saturating arithmetic.
  46. * The Q15 value -1 (0x8000) will be saturated to the maximum allowable positive value 0x7FFF.
  47. */
  48. void arm_abs_q15(
  49. q15_t * pSrc,
  50. q15_t * pDst,
  51. uint32_t blockSize)
  52. {
  53. uint32_t blkCnt; /* loop counter */
  54. #if defined (ARM_MATH_DSP)
  55. __SIMD32_TYPE *simd;
  56. /* Run the below code for Cortex-M4 and Cortex-M3 */
  57. q15_t in1; /* Input value1 */
  58. q15_t in2; /* Input value2 */
  59. /*loop Unrolling */
  60. blkCnt = blockSize >> 2U;
  61. /* First part of the processing with loop unrolling. Compute 4 outputs at a time.
  62. ** a second loop below computes the remaining 1 to 3 samples. */
  63. simd = __SIMD32_CONST(pDst);
  64. while (blkCnt > 0U)
  65. {
  66. /* C = |A| */
  67. /* Read two inputs */
  68. in1 = *pSrc++;
  69. in2 = *pSrc++;
  70. /* Store the Absolute result in the destination buffer by packing the two values, in a single cycle */
  71. #ifndef ARM_MATH_BIG_ENDIAN
  72. *simd++ =
  73. __PKHBT(((in1 > 0) ? in1 : (q15_t)__QSUB16(0, in1)),
  74. ((in2 > 0) ? in2 : (q15_t)__QSUB16(0, in2)), 16);
  75. #else
  76. *simd++ =
  77. __PKHBT(((in2 > 0) ? in2 : (q15_t)__QSUB16(0, in2)),
  78. ((in1 > 0) ? in1 : (q15_t)__QSUB16(0, in1)), 16);
  79. #endif /* #ifndef ARM_MATH_BIG_ENDIAN */
  80. in1 = *pSrc++;
  81. in2 = *pSrc++;
  82. #ifndef ARM_MATH_BIG_ENDIAN
  83. *simd++ =
  84. __PKHBT(((in1 > 0) ? in1 : (q15_t)__QSUB16(0, in1)),
  85. ((in2 > 0) ? in2 : (q15_t)__QSUB16(0, in2)), 16);
  86. #else
  87. *simd++ =
  88. __PKHBT(((in2 > 0) ? in2 : (q15_t)__QSUB16(0, in2)),
  89. ((in1 > 0) ? in1 : (q15_t)__QSUB16(0, in1)), 16);
  90. #endif /* #ifndef ARM_MATH_BIG_ENDIAN */
  91. /* Decrement the loop counter */
  92. blkCnt--;
  93. }
  94. pDst = (q15_t *)simd;
  95. /* If the blockSize is not a multiple of 4, compute any remaining output samples here.
  96. ** No loop unrolling is used. */
  97. blkCnt = blockSize % 0x4U;
  98. while (blkCnt > 0U)
  99. {
  100. /* C = |A| */
  101. /* Read the input */
  102. in1 = *pSrc++;
  103. /* Calculate absolute value of input and then store the result in the destination buffer. */
  104. *pDst++ = (in1 > 0) ? in1 : (q15_t)__QSUB16(0, in1);
  105. /* Decrement the loop counter */
  106. blkCnt--;
  107. }
  108. #else
  109. /* Run the below code for Cortex-M0 */
  110. q15_t in; /* Temporary input variable */
  111. /* Initialize blkCnt with number of samples */
  112. blkCnt = blockSize;
  113. while (blkCnt > 0U)
  114. {
  115. /* C = |A| */
  116. /* Read the input */
  117. in = *pSrc++;
  118. /* Calculate absolute value of input and then store the result in the destination buffer. */
  119. *pDst++ = (in > 0) ? in : ((in == (q15_t) 0x8000) ? 0x7fff : -in);
  120. /* Decrement the loop counter */
  121. blkCnt--;
  122. }
  123. #endif /* #if defined (ARM_MATH_DSP) */
  124. }
  125. /**
  126. * @} end of BasicAbs group
  127. */