arm_sub_q7.c 4.0 KB

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960616263646566676869707172737475767778798081828384858687888990919293949596979899100101102103104105106107108109110111112113114115116117118119120121122123124125126127128129130131132133134135136137138139140141142143144145146147148149150151152153154155156157
  1. /* ----------------------------------------------------------------------
  2. * Project: CMSIS DSP Library
  3. * Title: arm_sub_q7.c
  4. * Description: Q7 vector subtraction
  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 BasicSub
  34. @{
  35. */
  36. /**
  37. @brief Q7 vector subtraction.
  38. @param[in] pSrcA points to the first input vector
  39. @param[in] pSrcB points to the second input vector
  40. @param[out] pDst points to the output vector
  41. @param[in] blockSize number of samples in each vector
  42. @par Scaling and Overflow Behavior
  43. The function uses saturating arithmetic.
  44. Results outside of the allowable Q7 range [0x80 0x7F] will be saturated.
  45. */
  46. #if defined(ARM_MATH_MVEI) && !defined(ARM_MATH_AUTOVECTORIZE)
  47. #include "arm_helium_utils.h"
  48. void arm_sub_q7(
  49. const q7_t * pSrcA,
  50. const q7_t * pSrcB,
  51. q7_t * pDst,
  52. uint32_t blockSize)
  53. {
  54. uint32_t blkCnt; /* loop counters */
  55. q7x16_t vecA;
  56. q7x16_t vecB;
  57. /* Compute 16 outputs at a time */
  58. blkCnt = blockSize >> 4;
  59. while (blkCnt > 0U)
  60. {
  61. /*
  62. * C = A - B
  63. * Subtract and then store the results in the destination buffer.
  64. */
  65. vecA = vld1q(pSrcA);
  66. vecB = vld1q(pSrcB);
  67. vst1q(pDst, vqsubq(vecA, vecB));
  68. /*
  69. * Decrement the blockSize loop counter
  70. */
  71. blkCnt--;
  72. /*
  73. * advance vector source and destination pointers
  74. */
  75. pSrcA += 16;
  76. pSrcB += 16;
  77. pDst += 16;
  78. }
  79. /*
  80. * tail
  81. */
  82. blkCnt = blockSize & 0xF;
  83. if (blkCnt > 0U)
  84. {
  85. mve_pred16_t p0 = vctp8q(blkCnt);
  86. vecA = vld1q(pSrcA);
  87. vecB = vld1q(pSrcB);
  88. vstrbq_p(pDst, vqsubq(vecA, vecB), p0);
  89. }
  90. }
  91. #else
  92. void arm_sub_q7(
  93. const q7_t * pSrcA,
  94. const q7_t * pSrcB,
  95. q7_t * pDst,
  96. uint32_t blockSize)
  97. {
  98. uint32_t blkCnt; /* Loop counter */
  99. #if defined (ARM_MATH_LOOPUNROLL)
  100. /* Loop unrolling: Compute 4 outputs at a time */
  101. blkCnt = blockSize >> 2U;
  102. while (blkCnt > 0U)
  103. {
  104. /* C = A - B */
  105. #if defined (ARM_MATH_DSP)
  106. /* Subtract and store result in destination buffer (4 samples at a time). */
  107. write_q7x4_ia (&pDst, __QSUB8(read_q7x4_ia (&pSrcA), read_q7x4_ia (&pSrcB)));
  108. #else
  109. *pDst++ = (q7_t) __SSAT((q15_t) *pSrcA++ - *pSrcB++, 8);
  110. *pDst++ = (q7_t) __SSAT((q15_t) *pSrcA++ - *pSrcB++, 8);
  111. *pDst++ = (q7_t) __SSAT((q15_t) *pSrcA++ - *pSrcB++, 8);
  112. *pDst++ = (q7_t) __SSAT((q15_t) *pSrcA++ - *pSrcB++, 8);
  113. #endif
  114. /* Decrement loop counter */
  115. blkCnt--;
  116. }
  117. /* Loop unrolling: Compute remaining outputs */
  118. blkCnt = blockSize % 0x4U;
  119. #else
  120. /* Initialize blkCnt with number of samples */
  121. blkCnt = blockSize;
  122. #endif /* #if defined (ARM_MATH_LOOPUNROLL) */
  123. while (blkCnt > 0U)
  124. {
  125. /* C = A - B */
  126. /* Subtract and store result in destination buffer. */
  127. *pDst++ = (q7_t) __SSAT((q15_t) *pSrcA++ - *pSrcB++, 8);
  128. /* Decrement loop counter */
  129. blkCnt--;
  130. }
  131. }
  132. #endif /* defined(ARM_MATH_MVEI) */
  133. /**
  134. @} end of BasicSub group
  135. */