arm_offset_f32.c 4.8 KB

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960616263646566676869707172737475767778798081828384858687888990919293949596979899100101102103104105106107108109110111112113114115116117118119120121122123124125126127128129130131132133134135136137138139140141142143144145146147148149150151152153154155156157158159160161162163164165166167168169170171172173174175176177178179180181182183184185186187188189190191192193194195196
  1. /* ----------------------------------------------------------------------
  2. * Project: CMSIS DSP Library
  3. * Title: arm_offset_f32.c
  4. * Description: Floating-point vector offset
  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. @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. @return none
  53. */
  54. #if defined(ARM_MATH_MVEF) && !defined(ARM_MATH_AUTOVECTORIZE)
  55. #include "arm_helium_utils.h"
  56. void arm_offset_f32(
  57. const float32_t * pSrc,
  58. float32_t offset,
  59. float32_t * pDst,
  60. uint32_t blockSize)
  61. {
  62. uint32_t blkCnt; /* Loop counter */
  63. f32x4_t vec1;
  64. f32x4_t res;
  65. /* Compute 4 outputs at a time */
  66. blkCnt = blockSize >> 2U;
  67. while (blkCnt > 0U)
  68. {
  69. /* C = A + offset */
  70. /* Add offset and then store the results in the destination buffer. */
  71. vec1 = vld1q(pSrc);
  72. res = vaddq(vec1,offset);
  73. vst1q(pDst, res);
  74. /* Increment pointers */
  75. pSrc += 4;
  76. pDst += 4;
  77. /* Decrement the loop counter */
  78. blkCnt--;
  79. }
  80. /* Tail */
  81. blkCnt = blockSize & 0x3;
  82. if (blkCnt > 0U)
  83. {
  84. mve_pred16_t p0 = vctp32q(blkCnt);
  85. vec1 = vld1q((float32_t const *) pSrc);
  86. vstrwq_p(pDst, vaddq(vec1, offset), p0);
  87. }
  88. }
  89. #else
  90. void arm_offset_f32(
  91. const float32_t * pSrc,
  92. float32_t offset,
  93. float32_t * pDst,
  94. uint32_t blockSize)
  95. {
  96. uint32_t blkCnt; /* Loop counter */
  97. #if defined(ARM_MATH_NEON_EXPERIMENTAL) && !defined(ARM_MATH_AUTOVECTORIZE)
  98. f32x4_t vec1;
  99. f32x4_t res;
  100. /* Compute 4 outputs at a time */
  101. blkCnt = blockSize >> 2U;
  102. while (blkCnt > 0U)
  103. {
  104. /* C = A + offset */
  105. /* Add offset and then store the results in the destination buffer. */
  106. vec1 = vld1q_f32(pSrc);
  107. res = vaddq_f32(vec1,vdupq_n_f32(offset));
  108. vst1q_f32(pDst, res);
  109. /* Increment pointers */
  110. pSrc += 4;
  111. pDst += 4;
  112. /* Decrement the loop counter */
  113. blkCnt--;
  114. }
  115. /* Tail */
  116. blkCnt = blockSize & 0x3;
  117. #else
  118. #if defined (ARM_MATH_LOOPUNROLL) && !defined(ARM_MATH_AUTOVECTORIZE)
  119. /* Loop unrolling: Compute 4 outputs at a time */
  120. blkCnt = blockSize >> 2U;
  121. while (blkCnt > 0U)
  122. {
  123. /* C = A + offset */
  124. /* Add offset and store result in destination buffer. */
  125. *pDst++ = (*pSrc++) + offset;
  126. *pDst++ = (*pSrc++) + offset;
  127. *pDst++ = (*pSrc++) + offset;
  128. *pDst++ = (*pSrc++) + offset;
  129. /* Decrement loop counter */
  130. blkCnt--;
  131. }
  132. /* Loop unrolling: Compute remaining outputs */
  133. blkCnt = blockSize % 0x4U;
  134. #else
  135. /* Initialize blkCnt with number of samples */
  136. blkCnt = blockSize;
  137. #endif /* #if defined (ARM_MATH_LOOPUNROLL) */
  138. #endif /* #if defined(ARM_MATH_NEON_EXPERIMENTAL) */
  139. while (blkCnt > 0U)
  140. {
  141. /* C = A + offset */
  142. /* Add offset and store result in destination buffer. */
  143. *pDst++ = (*pSrc++) + offset;
  144. /* Decrement loop counter */
  145. blkCnt--;
  146. }
  147. }
  148. #endif /* defined(ARM_MATH_MVEF) && !defined(ARM_MATH_AUTOVECTORIZE) */
  149. /**
  150. @} end of BasicOffset group
  151. */