arm_abs_q15.c 4.6 KB

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960616263646566676869707172737475767778798081828384858687888990919293949596979899100101102103104105106107108109110111112113114115116117118119120121122123124125126127128129130131132133134135136137138139140141142143144145146147148149150151152153154155156157158159160161162163164165166167168169170171172173174175176177178
  1. /* ----------------------------------------------------------------------
  2. * Project: CMSIS DSP Library
  3. * Title: arm_abs_q15.c
  4. * Description: Q15 vector absolute value
  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 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. @return none
  42. @par Scaling and Overflow Behavior
  43. The function uses saturating arithmetic.
  44. The Q15 value -1 (0x8000) will be saturated to the maximum allowable positive value 0x7FFF.
  45. */
  46. #if defined(ARM_MATH_MVEI)
  47. #include "arm_helium_utils.h"
  48. void arm_abs_q15(
  49. const q15_t * pSrc,
  50. q15_t * pDst,
  51. uint32_t blockSize)
  52. {
  53. uint32_t blkCnt; /* loop counters */
  54. q15x8_t vecSrc;
  55. /* Compute 8 outputs at a time */
  56. blkCnt = blockSize >> 3;
  57. while (blkCnt > 0U)
  58. {
  59. /*
  60. * C = |A|
  61. * Calculate absolute and then store the results in the destination buffer.
  62. */
  63. vecSrc = vld1q(pSrc);
  64. vst1q(pDst, vqabsq(vecSrc));
  65. /*
  66. * Decrement the blockSize loop counter
  67. */
  68. blkCnt--;
  69. /*
  70. * advance vector source and destination pointers
  71. */
  72. pSrc += 8;
  73. pDst += 8;
  74. }
  75. /*
  76. * tail
  77. */
  78. blkCnt = blockSize & 7;
  79. if (blkCnt > 0U)
  80. {
  81. mve_pred16_t p0 = vctp16q(blkCnt);
  82. vecSrc = vld1q(pSrc);
  83. vstrhq_p(pDst, vqabsq(vecSrc), p0);
  84. }
  85. }
  86. #else
  87. void arm_abs_q15(
  88. const q15_t * pSrc,
  89. q15_t * pDst,
  90. uint32_t blockSize)
  91. {
  92. uint32_t blkCnt; /* Loop counter */
  93. q15_t in; /* Temporary input variable */
  94. #if defined (ARM_MATH_LOOPUNROLL)
  95. /* Loop unrolling: Compute 4 outputs at a time */
  96. blkCnt = blockSize >> 2U;
  97. while (blkCnt > 0U)
  98. {
  99. /* C = |A| */
  100. /* Calculate absolute of input (if -1 then saturated to 0x7fff) and store result in destination buffer. */
  101. in = *pSrc++;
  102. #if defined (ARM_MATH_DSP)
  103. *pDst++ = (in > 0) ? in : (q15_t)__QSUB16(0, in);
  104. #else
  105. *pDst++ = (in > 0) ? in : ((in == (q15_t) 0x8000) ? 0x7fff : -in);
  106. #endif
  107. in = *pSrc++;
  108. #if defined (ARM_MATH_DSP)
  109. *pDst++ = (in > 0) ? in : (q15_t)__QSUB16(0, in);
  110. #else
  111. *pDst++ = (in > 0) ? in : ((in == (q15_t) 0x8000) ? 0x7fff : -in);
  112. #endif
  113. in = *pSrc++;
  114. #if defined (ARM_MATH_DSP)
  115. *pDst++ = (in > 0) ? in : (q15_t)__QSUB16(0, in);
  116. #else
  117. *pDst++ = (in > 0) ? in : ((in == (q15_t) 0x8000) ? 0x7fff : -in);
  118. #endif
  119. in = *pSrc++;
  120. #if defined (ARM_MATH_DSP)
  121. *pDst++ = (in > 0) ? in : (q15_t)__QSUB16(0, in);
  122. #else
  123. *pDst++ = (in > 0) ? in : ((in == (q15_t) 0x8000) ? 0x7fff : -in);
  124. #endif
  125. /* Decrement loop counter */
  126. blkCnt--;
  127. }
  128. /* Loop unrolling: Compute remaining outputs */
  129. blkCnt = blockSize % 0x4U;
  130. #else
  131. /* Initialize blkCnt with number of samples */
  132. blkCnt = blockSize;
  133. #endif /* #if defined (ARM_MATH_LOOPUNROLL) */
  134. while (blkCnt > 0U)
  135. {
  136. /* C = |A| */
  137. /* Calculate absolute of input (if -1 then saturated to 0x7fff) and store result in destination buffer. */
  138. in = *pSrc++;
  139. #if defined (ARM_MATH_DSP)
  140. *pDst++ = (in > 0) ? in : (q15_t)__QSUB16(0, in);
  141. #else
  142. *pDst++ = (in > 0) ? in : ((in == (q15_t) 0x8000) ? 0x7fff : -in);
  143. #endif
  144. /* Decrement loop counter */
  145. blkCnt--;
  146. }
  147. }
  148. #endif /* defined(ARM_MATH_MVEI) */
  149. /**
  150. @} end of BasicAbs group
  151. */