arm_abs_q15.c 4.4 KB

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960616263646566676869707172737475767778798081828384858687888990919293949596979899100101102103104105106107108109110111112113114115116117118119120121122123124125126127128129130131132133134135136137138139140141142143144145146147148149150151152153154155156157158159160161162163164165166167168169170171172173174175176177
  1. /* ----------------------------------------------------------------------
  2. * Project: CMSIS DSP Library
  3. * Title: arm_abs_q15.c
  4. * Description: Q15 vector absolute value
  5. *
  6. * $Date: 23 April 2021
  7. * $Revision: V1.9.0
  8. *
  9. * Target Processor: Cortex-M and Cortex-A cores
  10. * -------------------------------------------------------------------- */
  11. /*
  12. * Copyright (C) 2010-2021 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 "dsp/basic_math_functions.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 vector
  39. @param[out] pDst points to the output vector
  40. @param[in] blockSize number of samples in each vector
  41. @par Scaling and Overflow Behavior
  42. The function uses saturating arithmetic.
  43. The Q15 value -1 (0x8000) will be saturated to the maximum allowable positive value 0x7FFF.
  44. */
  45. #if defined(ARM_MATH_MVEI) && !defined(ARM_MATH_AUTOVECTORIZE)
  46. #include "arm_helium_utils.h"
  47. void arm_abs_q15(
  48. const q15_t * pSrc,
  49. q15_t * pDst,
  50. uint32_t blockSize)
  51. {
  52. uint32_t blkCnt; /* loop counters */
  53. q15x8_t vecSrc;
  54. /* Compute 8 outputs at a time */
  55. blkCnt = blockSize >> 3;
  56. while (blkCnt > 0U)
  57. {
  58. /*
  59. * C = |A|
  60. * Calculate absolute and then store the results in the destination buffer.
  61. */
  62. vecSrc = vld1q(pSrc);
  63. vst1q(pDst, vqabsq(vecSrc));
  64. /*
  65. * Decrement the blockSize loop counter
  66. */
  67. blkCnt--;
  68. /*
  69. * advance vector source and destination pointers
  70. */
  71. pSrc += 8;
  72. pDst += 8;
  73. }
  74. /*
  75. * tail
  76. */
  77. blkCnt = blockSize & 7;
  78. if (blkCnt > 0U)
  79. {
  80. mve_pred16_t p0 = vctp16q(blkCnt);
  81. vecSrc = vld1q(pSrc);
  82. vstrhq_p(pDst, vqabsq(vecSrc), p0);
  83. }
  84. }
  85. #else
  86. void arm_abs_q15(
  87. const q15_t * pSrc,
  88. q15_t * pDst,
  89. uint32_t blockSize)
  90. {
  91. uint32_t blkCnt; /* Loop counter */
  92. q15_t in; /* Temporary input variable */
  93. #if defined (ARM_MATH_LOOPUNROLL)
  94. /* Loop unrolling: Compute 4 outputs at a time */
  95. blkCnt = blockSize >> 2U;
  96. while (blkCnt > 0U)
  97. {
  98. /* C = |A| */
  99. /* Calculate absolute of input (if -1 then saturated to 0x7fff) and store result in destination buffer. */
  100. in = *pSrc++;
  101. #if defined (ARM_MATH_DSP)
  102. *pDst++ = (in > 0) ? in : (q15_t)__QSUB16(0, in);
  103. #else
  104. *pDst++ = (in > 0) ? in : ((in == (q15_t) 0x8000) ? 0x7fff : -in);
  105. #endif
  106. in = *pSrc++;
  107. #if defined (ARM_MATH_DSP)
  108. *pDst++ = (in > 0) ? in : (q15_t)__QSUB16(0, in);
  109. #else
  110. *pDst++ = (in > 0) ? in : ((in == (q15_t) 0x8000) ? 0x7fff : -in);
  111. #endif
  112. in = *pSrc++;
  113. #if defined (ARM_MATH_DSP)
  114. *pDst++ = (in > 0) ? in : (q15_t)__QSUB16(0, in);
  115. #else
  116. *pDst++ = (in > 0) ? in : ((in == (q15_t) 0x8000) ? 0x7fff : -in);
  117. #endif
  118. in = *pSrc++;
  119. #if defined (ARM_MATH_DSP)
  120. *pDst++ = (in > 0) ? in : (q15_t)__QSUB16(0, in);
  121. #else
  122. *pDst++ = (in > 0) ? in : ((in == (q15_t) 0x8000) ? 0x7fff : -in);
  123. #endif
  124. /* Decrement loop counter */
  125. blkCnt--;
  126. }
  127. /* Loop unrolling: Compute remaining outputs */
  128. blkCnt = blockSize % 0x4U;
  129. #else
  130. /* Initialize blkCnt with number of samples */
  131. blkCnt = blockSize;
  132. #endif /* #if defined (ARM_MATH_LOOPUNROLL) */
  133. while (blkCnt > 0U)
  134. {
  135. /* C = |A| */
  136. /* Calculate absolute of input (if -1 then saturated to 0x7fff) and store result in destination buffer. */
  137. in = *pSrc++;
  138. #if defined (ARM_MATH_DSP)
  139. *pDst++ = (in > 0) ? in : (q15_t)__QSUB16(0, in);
  140. #else
  141. *pDst++ = (in > 0) ? in : ((in == (q15_t) 0x8000) ? 0x7fff : -in);
  142. #endif
  143. /* Decrement loop counter */
  144. blkCnt--;
  145. }
  146. }
  147. #endif /* defined(ARM_MATH_MVEI) */
  148. /**
  149. @} end of BasicAbs group
  150. */