Rev 2 | Details | Compare with Previous | Last modification | View Log | RSS feed
Rev | Author | Line No. | Line |
---|---|---|---|
2 | mjames | 1 | /* ---------------------------------------------------------------------- |
2 | * Project: CMSIS DSP Library |
||
3 | * Title: arm_var_f32.c |
||
4 | * Description: Variance of the elements of a floating-point vector |
||
5 | * |
||
6 | * $Date: 27. January 2017 |
||
7 | * $Revision: V.1.5.1 |
||
8 | * |
||
9 | * Target Processor: Cortex-M cores |
||
10 | * -------------------------------------------------------------------- */ |
||
11 | /* |
||
12 | * Copyright (C) 2010-2017 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 "arm_math.h" |
||
30 | |||
31 | /** |
||
32 | * @ingroup groupStats |
||
33 | */ |
||
34 | |||
35 | /** |
||
36 | * @defgroup variance Variance |
||
37 | * |
||
38 | * Calculates the variance of the elements in the input vector. |
||
39 | * The underlying algorithm used is the direct method sometimes referred to as the two-pass method: |
||
40 | * |
||
41 | * <pre> |
||
42 | * Result = sum(element - meanOfElements)^2) / numElement - 1 |
||
43 | * |
||
44 | * where, meanOfElements = ( pSrc[0] * pSrc[0] + pSrc[1] * pSrc[1] + ... + pSrc[blockSize-1] ) / blockSize |
||
45 | * |
||
46 | * </pre> |
||
47 | * |
||
48 | * There are separate functions for floating point, Q31, and Q15 data types. |
||
49 | */ |
||
50 | |||
51 | /** |
||
52 | * @addtogroup variance |
||
53 | * @{ |
||
54 | */ |
||
55 | |||
56 | |||
57 | /** |
||
58 | * @brief Variance of the elements of a floating-point vector. |
||
59 | * @param[in] *pSrc points to the input vector |
||
60 | * @param[in] blockSize length of the input vector |
||
61 | * @param[out] *pResult variance value returned here |
||
62 | * @return none. |
||
63 | */ |
||
64 | |||
65 | void arm_var_f32( |
||
66 | float32_t * pSrc, |
||
67 | uint32_t blockSize, |
||
68 | float32_t * pResult) |
||
69 | { |
||
70 | float32_t fMean, fValue; |
||
71 | uint32_t blkCnt; /* loop counter */ |
||
72 | float32_t * pInput = pSrc; |
||
73 | float32_t sum = 0.0f; |
||
74 | float32_t fSum = 0.0f; |
||
75 | #if defined(ARM_MATH_DSP) |
||
76 | float32_t in1, in2, in3, in4; |
||
77 | #endif |
||
78 | |||
79 | if (blockSize <= 1U) |
||
80 | { |
||
81 | *pResult = 0; |
||
82 | return; |
||
83 | } |
||
84 | |||
85 | #if defined(ARM_MATH_DSP) |
||
86 | /* Run the below code for Cortex-M4 and Cortex-M7 */ |
||
87 | |||
88 | /*loop Unrolling */ |
||
89 | blkCnt = blockSize >> 2U; |
||
90 | |||
91 | /* First part of the processing with loop unrolling. 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[1] + A[2] + ... + A[blockSize-1]) */ |
||
96 | in1 = *pInput++; |
||
97 | in2 = *pInput++; |
||
98 | in3 = *pInput++; |
||
99 | in4 = *pInput++; |
||
100 | |||
101 | sum += in1; |
||
102 | sum += in2; |
||
103 | sum += in3; |
||
104 | sum += in4; |
||
105 | |||
106 | /* Decrement the loop counter */ |
||
107 | blkCnt--; |
||
108 | } |
||
109 | |||
110 | /* If the blockSize is not a multiple of 4, compute any remaining output samples here. |
||
111 | ** No loop unrolling is used. */ |
||
112 | blkCnt = blockSize % 0x4U; |
||
113 | |||
114 | #else |
||
115 | /* Run the below code for Cortex-M0 or Cortex-M3 */ |
||
116 | |||
117 | /* Loop over blockSize number of values */ |
||
118 | blkCnt = blockSize; |
||
119 | |||
120 | #endif |
||
121 | |||
122 | while (blkCnt > 0U) |
||
123 | { |
||
124 | /* C = (A[0] + A[1] + A[2] + ... + A[blockSize-1]) */ |
||
125 | sum += *pInput++; |
||
126 | |||
127 | /* Decrement the loop counter */ |
||
128 | blkCnt--; |
||
129 | } |
||
130 | |||
131 | /* C = (A[0] + A[1] + A[2] + ... + A[blockSize-1]) / blockSize */ |
||
132 | fMean = sum / (float32_t) blockSize; |
||
133 | |||
134 | pInput = pSrc; |
||
135 | |||
136 | #if defined(ARM_MATH_DSP) |
||
137 | |||
138 | /*loop Unrolling */ |
||
139 | blkCnt = blockSize >> 2U; |
||
140 | |||
141 | /* First part of the processing with loop unrolling. Compute 4 outputs at a time. |
||
142 | ** a second loop below computes the remaining 1 to 3 samples. */ |
||
143 | while (blkCnt > 0U) |
||
144 | { |
||
145 | fValue = *pInput++ - fMean; |
||
146 | fSum += fValue * fValue; |
||
147 | fValue = *pInput++ - fMean; |
||
148 | fSum += fValue * fValue; |
||
149 | fValue = *pInput++ - fMean; |
||
150 | fSum += fValue * fValue; |
||
151 | fValue = *pInput++ - fMean; |
||
152 | fSum += fValue * fValue; |
||
153 | |||
154 | /* Decrement the loop counter */ |
||
155 | blkCnt--; |
||
156 | } |
||
157 | |||
158 | blkCnt = blockSize % 0x4U; |
||
159 | #else |
||
160 | /* Run the below code for Cortex-M0 or Cortex-M3 */ |
||
161 | |||
162 | /* Loop over blockSize number of values */ |
||
163 | blkCnt = blockSize; |
||
164 | #endif |
||
165 | |||
166 | while (blkCnt > 0U) |
||
167 | { |
||
168 | fValue = *pInput++ - fMean; |
||
169 | fSum += fValue * fValue; |
||
170 | |||
171 | /* Decrement the loop counter */ |
||
172 | blkCnt--; |
||
173 | } |
||
174 | |||
175 | /* Variance */ |
||
176 | *pResult = fSum / (float32_t)(blockSize - 1.0f); |
||
177 | } |
||
178 | |||
179 | /** |
||
180 | * @} end of variance group |
||
181 | */ |