1 /* ----------------------------------------------------------------------
2 * Project: CMSIS DSP Library
3 * Title: arm_rms_f32.c
4 * Description: Root mean square value of the elements of a floating-point vector
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
29 #include "dsp/statistics_functions.h"
30
31 /**
32 @ingroup groupStats
33 */
34
35 /**
36 @defgroup RMS Root mean square (RMS)
37
38 Calculates the Root Mean Square of the elements in the input vector.
39 The underlying algorithm is used:
40
41 <pre>
42 Result = sqrt(((pSrc[0] * pSrc[0] + pSrc[1] * pSrc[1] + ... + pSrc[blockSize-1] * pSrc[blockSize-1]) / blockSize));
43 </pre>
44
45 There are separate functions for floating point, Q31, and Q15 data types.
46 */
47
48 /**
49 @addtogroup RMS
50 @{
51 */
52
53 /**
54 @brief Root Mean Square of the elements of a floating-point vector.
55 @param[in] pSrc points to the input vector
56 @param[in] blockSize number of samples in input vector
57 @param[out] pResult root mean square value returned here
58 @return none
59 */
60
61 #if defined(ARM_MATH_MVEF) && !defined(ARM_MATH_AUTOVECTORIZE)
arm_rms_f32(const float32_t * pSrc,uint32_t blockSize,float32_t * pResult)62 void arm_rms_f32(
63 const float32_t * pSrc,
64 uint32_t blockSize,
65 float32_t * pResult)
66 {
67 float32_t pow = 0.0f;
68
69 arm_power_f32(pSrc, blockSize, &pow);
70
71 /* Compute Rms and store the result in the destination */
72 arm_sqrt_f32(pow / (float32_t) blockSize, pResult);
73 }
74 #else
75 #if defined(ARM_MATH_NEON) && !defined(ARM_MATH_AUTOVECTORIZE)
arm_rms_f32(const float32_t * pSrc,uint32_t blockSize,float32_t * pResult)76 void arm_rms_f32(
77 const float32_t * pSrc,
78 uint32_t blockSize,
79 float32_t * pResult)
80 {
81 float32_t sum = 0.0f; /* accumulator */
82 float32_t in; /* Temporary variable to store input value */
83 uint32_t blkCnt; /* loop counter */
84
85 float32x4_t sumV = vdupq_n_f32(0.0f); /* Temporary result storage */
86 float32x2_t sumV2;
87 float32x4_t inV;
88
89 blkCnt = blockSize >> 2U;
90
91 /* Compute 4 outputs at a time.
92 ** a second loop below computes the remaining 1 to 3 samples. */
93 while (blkCnt > 0U)
94 {
95 /* C = A[0] * A[0] + A[1] * A[1] + A[2] * A[2] + ... + A[blockSize-1] * A[blockSize-1] */
96 /* Compute Power and then store the result in a temporary variable, sum. */
97 inV = vld1q_f32(pSrc);
98 sumV = vmlaq_f32(sumV, inV, inV);
99 pSrc += 4;
100
101 /* Decrement the loop counter */
102 blkCnt--;
103 }
104
105 sumV2 = vpadd_f32(vget_low_f32(sumV),vget_high_f32(sumV));
106 sum = vget_lane_f32(sumV2, 0) + vget_lane_f32(sumV2, 1);
107
108 /* If the blockSize is not a multiple of 4, compute any remaining output samples here.
109 ** No loop unrolling is used. */
110 blkCnt = blockSize % 0x4U;
111
112 while (blkCnt > 0U)
113 {
114 /* C = A[0] * A[0] + A[1] * A[1] + A[2] * A[2] + ... + A[blockSize-1] * A[blockSize-1] */
115 /* compute power and then store the result in a temporary variable, sum. */
116 in = *pSrc++;
117 sum += in * in;
118
119 /* Decrement the loop counter */
120 blkCnt--;
121 }
122
123 /* Compute Rms and store the result in the destination */
124 arm_sqrt_f32(sum / (float32_t) blockSize, pResult);
125 }
126 #else
arm_rms_f32(const float32_t * pSrc,uint32_t blockSize,float32_t * pResult)127 void arm_rms_f32(
128 const float32_t * pSrc,
129 uint32_t blockSize,
130 float32_t * pResult)
131 {
132 uint32_t blkCnt; /* Loop counter */
133 float32_t sum = 0.0f; /* Temporary result storage */
134 float32_t in; /* Temporary variable to store input value */
135
136 #if defined (ARM_MATH_LOOPUNROLL) && !defined(ARM_MATH_AUTOVECTORIZE)
137
138 /* Loop unrolling: Compute 4 outputs at a time */
139 blkCnt = blockSize >> 2U;
140
141 while (blkCnt > 0U)
142 {
143 /* C = A[0] * A[0] + A[1] * A[1] + ... + A[blockSize-1] * A[blockSize-1] */
144
145 in = *pSrc++;
146 /* Compute sum of squares and store result in a temporary variable, sum. */
147 sum += in * in;
148
149 in = *pSrc++;
150 sum += in * in;
151
152 in = *pSrc++;
153 sum += in * in;
154
155 in = *pSrc++;
156 sum += in * in;
157
158 /* Decrement loop counter */
159 blkCnt--;
160 }
161
162 /* Loop unrolling: Compute remaining outputs */
163 blkCnt = blockSize % 0x4U;
164
165 #else
166
167 /* Initialize blkCnt with number of samples */
168 blkCnt = blockSize;
169
170 #endif /* #if defined (ARM_MATH_LOOPUNROLL) */
171
172 while (blkCnt > 0U)
173 {
174 /* C = A[0] * A[0] + A[1] * A[1] + ... + A[blockSize-1] * A[blockSize-1] */
175
176 in = *pSrc++;
177 /* Compute sum of squares and store result in a temporary variable. */
178 sum += ( in * in);
179
180 /* Decrement loop counter */
181 blkCnt--;
182 }
183
184 /* Compute Rms and store result in destination */
185 arm_sqrt_f32(sum / (float32_t) blockSize, pResult);
186 }
187 #endif /* #if defined(ARM_MATH_NEON) */
188 #endif /* defined(ARM_MATH_MVEF) && !defined(ARM_MATH_AUTOVECTORIZE) */
189
190 /**
191 @} end of RMS group
192 */
193