arm_abs_q7.c 4.5 KB

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960616263646566676869707172737475767778798081828384858687888990919293949596979899100101102103104105106107108109110111112113114115116117118119120121122123124125126127128129130131132133134135136137138139140141142143144145146147148149150151152153154155156157158159160161162163164165166167168169170171172173174175176177178179
  1. /* ----------------------------------------------------------------------
  2. * Project: CMSIS DSP Library
  3. * Title: arm_abs_q7.c
  4. * Description: Q7 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 Q7 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 Conditions for optimum performance
  42. Input and output buffers should be aligned by 32-bit
  43. @par Scaling and Overflow Behavior
  44. The function uses saturating arithmetic.
  45. The Q7 value -1 (0x80) will be saturated to the maximum allowable positive value 0x7F.
  46. */
  47. #if defined(ARM_MATH_MVEI) && !defined(ARM_MATH_AUTOVECTORIZE)
  48. #include "arm_helium_utils.h"
  49. void arm_abs_q7(
  50. const q7_t * pSrc,
  51. q7_t * pDst,
  52. uint32_t blockSize)
  53. {
  54. uint32_t blkCnt; /* loop counters */
  55. q7x16_t vecSrc;
  56. /* Compute 16 outputs at a time */
  57. blkCnt = blockSize >> 4;
  58. while (blkCnt > 0U)
  59. {
  60. /*
  61. * C = |A|
  62. * Calculate absolute and then store the results in the destination buffer.
  63. */
  64. vecSrc = vld1q(pSrc);
  65. vst1q(pDst, vqabsq(vecSrc));
  66. /*
  67. * Decrement the blockSize loop counter
  68. */
  69. blkCnt--;
  70. /*
  71. * advance vector source and destination pointers
  72. */
  73. pSrc += 16;
  74. pDst += 16;
  75. }
  76. /*
  77. * tail
  78. */
  79. blkCnt = blockSize & 0xF;
  80. if (blkCnt > 0U)
  81. {
  82. mve_pred16_t p0 = vctp8q(blkCnt);
  83. vecSrc = vld1q(pSrc);
  84. vstrbq_p(pDst, vqabsq(vecSrc), p0);
  85. }
  86. }
  87. #else
  88. void arm_abs_q7(
  89. const q7_t * pSrc,
  90. q7_t * pDst,
  91. uint32_t blockSize)
  92. {
  93. uint32_t blkCnt; /* Loop counter */
  94. q7_t in; /* Temporary input variable */
  95. #if defined (ARM_MATH_LOOPUNROLL)
  96. /* Loop unrolling: Compute 4 outputs at a time */
  97. blkCnt = blockSize >> 2U;
  98. while (blkCnt > 0U)
  99. {
  100. /* C = |A| */
  101. /* Calculate absolute of input (if -1 then saturated to 0x7f) and store result in destination buffer. */
  102. in = *pSrc++;
  103. #if defined (ARM_MATH_DSP)
  104. *pDst++ = (in > 0) ? in : (q7_t)__QSUB8(0, in);
  105. #else
  106. *pDst++ = (in > 0) ? in : ((in == (q7_t) 0x80) ? (q7_t) 0x7f : -in);
  107. #endif
  108. in = *pSrc++;
  109. #if defined (ARM_MATH_DSP)
  110. *pDst++ = (in > 0) ? in : (q7_t)__QSUB8(0, in);
  111. #else
  112. *pDst++ = (in > 0) ? in : ((in == (q7_t) 0x80) ? (q7_t) 0x7f : -in);
  113. #endif
  114. in = *pSrc++;
  115. #if defined (ARM_MATH_DSP)
  116. *pDst++ = (in > 0) ? in : (q7_t)__QSUB8(0, in);
  117. #else
  118. *pDst++ = (in > 0) ? in : ((in == (q7_t) 0x80) ? (q7_t) 0x7f : -in);
  119. #endif
  120. in = *pSrc++;
  121. #if defined (ARM_MATH_DSP)
  122. *pDst++ = (in > 0) ? in : (q7_t)__QSUB8(0, in);
  123. #else
  124. *pDst++ = (in > 0) ? in : ((in == (q7_t) 0x80) ? (q7_t) 0x7f : -in);
  125. #endif
  126. /* Decrement loop counter */
  127. blkCnt--;
  128. }
  129. /* Loop unrolling: Compute remaining outputs */
  130. blkCnt = blockSize % 0x4U;
  131. #else
  132. /* Initialize blkCnt with number of samples */
  133. blkCnt = blockSize;
  134. #endif /* #if defined (ARM_MATH_LOOPUNROLL) */
  135. while (blkCnt > 0U)
  136. {
  137. /* C = |A| */
  138. /* Calculate absolute of input (if -1 then saturated to 0x7f) and store result in destination buffer. */
  139. in = *pSrc++;
  140. #if defined (ARM_MATH_DSP)
  141. *pDst++ = (in > 0) ? in : (q7_t) __QSUB8(0, in);
  142. #else
  143. *pDst++ = (in > 0) ? in : ((in == (q7_t) 0x80) ? (q7_t) 0x7f : -in);
  144. #endif
  145. /* Decrement loop counter */
  146. blkCnt--;
  147. }
  148. }
  149. #endif /* defined(ARM_MATH_MVEI) */
  150. /**
  151. @} end of BasicAbs group
  152. */