arm_offset_f16.c 3.9 KB

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960616263646566676869707172737475767778798081828384858687888990919293949596979899100101102103104105106107108109110111112113114115116117118119120121122123124125126127128129130131132133134135136137138139140141142143144145146147148149150151152153154155156157158159160161162163164165166167168
  1. /* ----------------------------------------------------------------------
  2. * Project: CMSIS DSP Library
  3. * Title: arm_offset_f16.c
  4. * Description: Floating-point vector offset
  5. *
  6. *
  7. * Target Processor: Cortex-M cores
  8. * -------------------------------------------------------------------- */
  9. /*
  10. * Copyright (C) 2010-2020 ARM Limited or its affiliates. All rights reserved.
  11. *
  12. * SPDX-License-Identifier: Apache-2.0
  13. *
  14. * Licensed under the Apache License, Version 2.0 (the License); you may
  15. * not use this file except in compliance with the License.
  16. * You may obtain a copy of the License at
  17. *
  18. * www.apache.org/licenses/LICENSE-2.0
  19. *
  20. * Unless required by applicable law or agreed to in writing, software
  21. * distributed under the License is distributed on an AS IS BASIS, WITHOUT
  22. * WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
  23. * See the License for the specific language governing permissions and
  24. * limitations under the License.
  25. */
  26. #include "dsp/basic_math_functions_f16.h"
  27. /**
  28. @ingroup groupMath
  29. */
  30. /**
  31. @defgroup BasicOffset Vector Offset
  32. Adds a constant offset to each element of a vector.
  33. <pre>
  34. pDst[n] = pSrc[n] + offset, 0 <= n < blockSize.
  35. </pre>
  36. The functions support in-place computation allowing the source and
  37. destination pointers to reference the same memory buffer.
  38. There are separate functions for floating-point, Q7, Q15, and Q31 data types.
  39. */
  40. /**
  41. @addtogroup BasicOffset
  42. @{
  43. */
  44. /**
  45. @brief Adds a constant offset to a floating-point vector.
  46. @param[in] pSrc points to the input vector
  47. @param[in] offset is the offset to be added
  48. @param[out] pDst points to the output vector
  49. @param[in] blockSize number of samples in each vector
  50. @return none
  51. */
  52. #if defined(ARM_MATH_MVE_FLOAT16) && !defined(ARM_MATH_AUTOVECTORIZE)
  53. #include "arm_helium_utils.h"
  54. void arm_offset_f16(
  55. const float16_t * pSrc,
  56. float16_t offset,
  57. float16_t * pDst,
  58. uint32_t blockSize)
  59. {
  60. uint32_t blkCnt; /* Loop counter */
  61. f16x8_t vec1;
  62. f16x8_t res;
  63. /* Compute 4 outputs at a time */
  64. blkCnt = blockSize >> 3U;
  65. while (blkCnt > 0U)
  66. {
  67. /* C = A + offset */
  68. /* Add offset and then store the results in the destination buffer. */
  69. vec1 = vld1q(pSrc);
  70. res = vaddq(vec1,offset);
  71. vst1q(pDst, res);
  72. /* Increment pointers */
  73. pSrc += 8;
  74. pDst += 8;
  75. /* Decrement the loop counter */
  76. blkCnt--;
  77. }
  78. /* Tail */
  79. blkCnt = blockSize & 0x7;
  80. if (blkCnt > 0U)
  81. {
  82. mve_pred16_t p0 = vctp16q(blkCnt);
  83. vec1 = vld1q((float16_t const *) pSrc);
  84. vstrhq_p(pDst, vaddq(vec1, offset), p0);
  85. }
  86. }
  87. #else
  88. #if defined(ARM_FLOAT16_SUPPORTED)
  89. void arm_offset_f16(
  90. const float16_t * pSrc,
  91. float16_t offset,
  92. float16_t * pDst,
  93. uint32_t blockSize)
  94. {
  95. uint32_t blkCnt; /* Loop counter */
  96. #if defined (ARM_MATH_LOOPUNROLL) && !defined(ARM_MATH_AUTOVECTORIZE)
  97. /* Loop unrolling: Compute 4 outputs at a time */
  98. blkCnt = blockSize >> 2U;
  99. while (blkCnt > 0U)
  100. {
  101. /* C = A + offset */
  102. /* Add offset and store result in destination buffer. */
  103. *pDst++ = (*pSrc++) + offset;
  104. *pDst++ = (*pSrc++) + offset;
  105. *pDst++ = (*pSrc++) + offset;
  106. *pDst++ = (*pSrc++) + offset;
  107. /* Decrement loop counter */
  108. blkCnt--;
  109. }
  110. /* Loop unrolling: Compute remaining outputs */
  111. blkCnt = blockSize % 0x4U;
  112. #else
  113. /* Initialize blkCnt with number of samples */
  114. blkCnt = blockSize;
  115. #endif /* #if defined (ARM_MATH_LOOPUNROLL) */
  116. while (blkCnt > 0U)
  117. {
  118. /* C = A + offset */
  119. /* Add offset and store result in destination buffer. */
  120. *pDst++ = (*pSrc++) + offset;
  121. /* Decrement loop counter */
  122. blkCnt--;
  123. }
  124. }
  125. #endif
  126. #endif /* defined(ARM_MATH_MVEF) && !defined(ARM_MATH_AUTOVECTORIZE) */
  127. /**
  128. @} end of BasicOffset group
  129. */