arm_power_f32.c 4.7 KB

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960616263646566676869707172737475767778798081828384858687888990919293949596979899100101102103104105106107108109110111112113114115116117118119120121122123124125126127128129130131132133134135136137138139140141142143144145146147148149150151152153154155156157158159160161162163164165166167168169170171172173174175
  1. /* ----------------------------------------------------------------------
  2. * Project: CMSIS DSP Library
  3. * Title: arm_power_f32.c
  4. * Description: Sum of the squares of the elements of a floating-point vector
  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 groupStats
  31. */
  32. /**
  33. @defgroup power Power
  34. Calculates the sum of the squares of the elements in the input vector.
  35. The underlying algorithm is used:
  36. <pre>
  37. Result = pSrc[0] * pSrc[0] + pSrc[1] * pSrc[1] + pSrc[2] * pSrc[2] + ... + pSrc[blockSize-1] * pSrc[blockSize-1];
  38. </pre>
  39. There are separate functions for floating point, Q31, Q15, and Q7 data types.
  40. */
  41. /**
  42. @addtogroup power
  43. @{
  44. */
  45. /**
  46. @brief Sum of the squares of the elements of a floating-point vector.
  47. @param[in] pSrc points to the input vector
  48. @param[in] blockSize number of samples in input vector
  49. @param[out] pResult sum of the squares value returned here
  50. @return none
  51. */
  52. #if defined(ARM_MATH_NEON)
  53. void arm_power_f32(
  54. const float32_t * pSrc,
  55. uint32_t blockSize,
  56. float32_t * pResult)
  57. {
  58. float32_t sum = 0.0f; /* accumulator */
  59. float32_t in; /* Temporary variable to store input value */
  60. uint32_t blkCnt; /* loop counter */
  61. float32x4_t sumV = vdupq_n_f32(0.0f); /* Temporary result storage */
  62. float32x2_t sumV2;
  63. float32x4_t inV;
  64. blkCnt = blockSize >> 2U;
  65. /* Compute 4 outputs at a time.
  66. ** a second loop below computes the remaining 1 to 3 samples. */
  67. while (blkCnt > 0U)
  68. {
  69. /* C = A[0] * A[0] + A[1] * A[1] + A[2] * A[2] + ... + A[blockSize-1] * A[blockSize-1] */
  70. /* Compute Power and then store the result in a temporary variable, sum. */
  71. inV = vld1q_f32(pSrc);
  72. sumV = vmlaq_f32(sumV, inV, inV);
  73. pSrc += 4;
  74. /* Decrement the loop counter */
  75. blkCnt--;
  76. }
  77. sumV2 = vpadd_f32(vget_low_f32(sumV),vget_high_f32(sumV));
  78. sum = sumV2[0] + sumV2[1];
  79. /* If the blockSize is not a multiple of 4, compute any remaining output samples here.
  80. ** No loop unrolling is used. */
  81. blkCnt = blockSize % 0x4U;
  82. while (blkCnt > 0U)
  83. {
  84. /* C = A[0] * A[0] + A[1] * A[1] + A[2] * A[2] + ... + A[blockSize-1] * A[blockSize-1] */
  85. /* compute power and then store the result in a temporary variable, sum. */
  86. in = *pSrc++;
  87. sum += in * in;
  88. /* Decrement the loop counter */
  89. blkCnt--;
  90. }
  91. /* Store the result to the destination */
  92. *pResult = sum;
  93. }
  94. #else
  95. void arm_power_f32(
  96. const float32_t * pSrc,
  97. uint32_t blockSize,
  98. float32_t * pResult)
  99. {
  100. uint32_t blkCnt; /* Loop counter */
  101. float32_t sum = 0.0f; /* Temporary result storage */
  102. float32_t in; /* Temporary variable to store input value */
  103. #if defined (ARM_MATH_LOOPUNROLL)
  104. /* Loop unrolling: Compute 4 outputs at a time */
  105. blkCnt = blockSize >> 2U;
  106. while (blkCnt > 0U)
  107. {
  108. /* C = A[0] * A[0] + A[1] * A[1] + ... + A[blockSize-1] * A[blockSize-1] */
  109. /* Compute Power and store result in a temporary variable, sum. */
  110. in = *pSrc++;
  111. sum += in * in;
  112. in = *pSrc++;
  113. sum += in * in;
  114. in = *pSrc++;
  115. sum += in * in;
  116. in = *pSrc++;
  117. sum += in * in;
  118. /* Decrement loop counter */
  119. blkCnt--;
  120. }
  121. /* Loop unrolling: Compute remaining outputs */
  122. blkCnt = blockSize % 0x4U;
  123. #else
  124. /* Initialize blkCnt with number of samples */
  125. blkCnt = blockSize;
  126. #endif /* #if defined (ARM_MATH_LOOPUNROLL) */
  127. while (blkCnt > 0U)
  128. {
  129. /* C = A[0] * A[0] + A[1] * A[1] + ... + A[blockSize-1] * A[blockSize-1] */
  130. /* Compute Power and store result in a temporary variable, sum. */
  131. in = *pSrc++;
  132. sum += in * in;
  133. /* Decrement loop counter */
  134. blkCnt--;
  135. }
  136. /* Store result to destination */
  137. *pResult = sum;
  138. }
  139. #endif /* #if defined(ARM_MATH_NEON) */
  140. /**
  141. @} end of power group
  142. */