arm_offset_f32.c 4.6 KB

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960616263646566676869707172737475767778798081828384858687888990919293949596979899100101102103104105106107108109110111112113114115116117118119120121122123124125126127128129130131132133134135136137138139140141142143144145146147148149150151152153154155156157158159160161162163164165166167168169170171172173174175176177178179180181182183184185186187188189190191192193194195
  1. /* ----------------------------------------------------------------------
  2. * Project: CMSIS DSP Library
  3. * Title: arm_offset_f32.c
  4. * Description: Floating-point vector offset
  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. @defgroup BasicOffset Vector Offset
  34. Adds a constant offset to each element of a vector.
  35. <pre>
  36. pDst[n] = pSrc[n] + offset, 0 <= n < blockSize.
  37. </pre>
  38. The functions support in-place computation allowing the source and
  39. destination pointers to reference the same memory buffer.
  40. There are separate functions for floating-point, Q7, Q15, and Q31 data types.
  41. */
  42. /**
  43. @addtogroup BasicOffset
  44. @{
  45. */
  46. /**
  47. @brief Adds a constant offset to a floating-point vector.
  48. @param[in] pSrc points to the input vector
  49. @param[in] offset is the offset to be added
  50. @param[out] pDst points to the output vector
  51. @param[in] blockSize number of samples in each vector
  52. */
  53. #if defined(ARM_MATH_MVEF) && !defined(ARM_MATH_AUTOVECTORIZE)
  54. #include "arm_helium_utils.h"
  55. void arm_offset_f32(
  56. const float32_t * pSrc,
  57. float32_t offset,
  58. float32_t * pDst,
  59. uint32_t blockSize)
  60. {
  61. uint32_t blkCnt; /* Loop counter */
  62. f32x4_t vec1;
  63. f32x4_t res;
  64. /* Compute 4 outputs at a time */
  65. blkCnt = blockSize >> 2U;
  66. while (blkCnt > 0U)
  67. {
  68. /* C = A + offset */
  69. /* Add offset and then store the results in the destination buffer. */
  70. vec1 = vld1q(pSrc);
  71. res = vaddq(vec1,offset);
  72. vst1q(pDst, res);
  73. /* Increment pointers */
  74. pSrc += 4;
  75. pDst += 4;
  76. /* Decrement the loop counter */
  77. blkCnt--;
  78. }
  79. /* Tail */
  80. blkCnt = blockSize & 0x3;
  81. if (blkCnt > 0U)
  82. {
  83. mve_pred16_t p0 = vctp32q(blkCnt);
  84. vec1 = vld1q((float32_t const *) pSrc);
  85. vstrwq_p(pDst, vaddq(vec1, offset), p0);
  86. }
  87. }
  88. #else
  89. void arm_offset_f32(
  90. const float32_t * pSrc,
  91. float32_t offset,
  92. float32_t * pDst,
  93. uint32_t blockSize)
  94. {
  95. uint32_t blkCnt; /* Loop counter */
  96. #if defined(ARM_MATH_NEON_EXPERIMENTAL) && !defined(ARM_MATH_AUTOVECTORIZE)
  97. f32x4_t vec1;
  98. f32x4_t res;
  99. /* Compute 4 outputs at a time */
  100. blkCnt = blockSize >> 2U;
  101. while (blkCnt > 0U)
  102. {
  103. /* C = A + offset */
  104. /* Add offset and then store the results in the destination buffer. */
  105. vec1 = vld1q_f32(pSrc);
  106. res = vaddq_f32(vec1,vdupq_n_f32(offset));
  107. vst1q_f32(pDst, res);
  108. /* Increment pointers */
  109. pSrc += 4;
  110. pDst += 4;
  111. /* Decrement the loop counter */
  112. blkCnt--;
  113. }
  114. /* Tail */
  115. blkCnt = blockSize & 0x3;
  116. #else
  117. #if defined (ARM_MATH_LOOPUNROLL) && !defined(ARM_MATH_AUTOVECTORIZE)
  118. /* Loop unrolling: Compute 4 outputs at a time */
  119. blkCnt = blockSize >> 2U;
  120. while (blkCnt > 0U)
  121. {
  122. /* C = A + offset */
  123. /* Add offset and store result in destination buffer. */
  124. *pDst++ = (*pSrc++) + offset;
  125. *pDst++ = (*pSrc++) + offset;
  126. *pDst++ = (*pSrc++) + offset;
  127. *pDst++ = (*pSrc++) + offset;
  128. /* Decrement loop counter */
  129. blkCnt--;
  130. }
  131. /* Loop unrolling: Compute remaining outputs */
  132. blkCnt = blockSize % 0x4U;
  133. #else
  134. /* Initialize blkCnt with number of samples */
  135. blkCnt = blockSize;
  136. #endif /* #if defined (ARM_MATH_LOOPUNROLL) */
  137. #endif /* #if defined(ARM_MATH_NEON_EXPERIMENTAL) */
  138. while (blkCnt > 0U)
  139. {
  140. /* C = A + offset */
  141. /* Add offset and store result in destination buffer. */
  142. *pDst++ = (*pSrc++) + offset;
  143. /* Decrement loop counter */
  144. blkCnt--;
  145. }
  146. }
  147. #endif /* defined(ARM_MATH_MVEF) && !defined(ARM_MATH_AUTOVECTORIZE) */
  148. /**
  149. @} end of BasicOffset group
  150. */