1 /* ----------------------------------------------------------------------
2 * Project: CMSIS DSP Library
3 * Title: arm_mean_q15.c
4 * Description: Mean value of a Q15 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 @addtogroup mean
37 @{
38 */
39
40 /**
41 @brief Mean value of a Q15 vector.
42 @param[in] pSrc points to the input vector
43 @param[in] blockSize number of samples in input vector
44 @param[out] pResult mean value returned here
45 @return none
46
47 @par Scaling and Overflow Behavior
48 The function is implemented using a 32-bit internal accumulator.
49 The input is represented in 1.15 format and is accumulated in a 32-bit
50 accumulator in 17.15 format.
51 There is no risk of internal overflow with this approach, and the
52 full precision of intermediate result is preserved.
53 Finally, the accumulator is truncated to yield a result of 1.15 format.
54 */
55
56 #if defined(ARM_MATH_MVEI) && !defined(ARM_MATH_AUTOVECTORIZE)
arm_mean_q15(const q15_t * pSrc,uint32_t blockSize,q15_t * pResult)57 void arm_mean_q15(
58 const q15_t * pSrc,
59 uint32_t blockSize,
60 q15_t * pResult)
61 {
62 uint32_t blkCnt; /* loop counters */
63 q15x8_t vecSrc;
64 q31_t sum = 0L;
65
66 /* Compute 8 outputs at a time */
67 blkCnt = blockSize >> 3U;
68 while (blkCnt > 0U)
69 {
70 vecSrc = vldrhq_s16(pSrc);
71 /*
72 * sum lanes
73 */
74 sum = vaddvaq(sum, vecSrc);
75
76 blkCnt--;
77 pSrc += 8;
78 }
79
80 /* Tail */
81 blkCnt = blockSize & 0x7;
82
83 while (blkCnt > 0U)
84 {
85 /* C = (A[0] + A[1] + A[2] + ... + A[blockSize-1]) */
86 sum += *pSrc++;
87
88 /* Decrement loop counter */
89 blkCnt--;
90 }
91
92 /* C = (A[0] + A[1] + A[2] + ... + A[blockSize-1]) / blockSize */
93 /* Store the result to the destination */
94 *pResult = (q15_t) (sum / (int32_t) blockSize);
95 }
96 #else
arm_mean_q15(const q15_t * pSrc,uint32_t blockSize,q15_t * pResult)97 void arm_mean_q15(
98 const q15_t * pSrc,
99 uint32_t blockSize,
100 q15_t * pResult)
101 {
102 uint32_t blkCnt; /* Loop counter */
103 q31_t sum = 0; /* Temporary result storage */
104
105 #if defined (ARM_MATH_LOOPUNROLL)
106 q31_t in;
107 #endif
108
109 #if defined (ARM_MATH_LOOPUNROLL)
110
111 /* Loop unrolling: Compute 4 outputs at a time */
112 blkCnt = blockSize >> 2U;
113
114 while (blkCnt > 0U)
115 {
116 /* C = (A[0] + A[1] + A[2] + ... + A[blockSize-1]) */
117 in = read_q15x2_ia ((q15_t **) &pSrc);
118 sum += ((in << 16U) >> 16U);
119 sum += (in >> 16U);
120
121 in = read_q15x2_ia ((q15_t **) &pSrc);
122 sum += ((in << 16U) >> 16U);
123 sum += (in >> 16U);
124
125 /* Decrement the loop counter */
126 blkCnt--;
127 }
128
129 /* Loop unrolling: Compute remaining outputs */
130 blkCnt = blockSize % 0x4U;
131
132 #else
133
134 /* Initialize blkCnt with number of samples */
135 blkCnt = blockSize;
136
137 #endif /* #if defined (ARM_MATH_LOOPUNROLL) */
138
139 while (blkCnt > 0U)
140 {
141 /* C = (A[0] + A[1] + A[2] + ... + A[blockSize-1]) */
142 sum += *pSrc++;
143
144 /* Decrement loop counter */
145 blkCnt--;
146 }
147
148 /* C = (A[0] + A[1] + A[2] + ... + A[blockSize-1]) / blockSize */
149 /* Store result to destination */
150 *pResult = (q15_t) (sum / (int32_t) blockSize);
151 }
152 #endif /* defined(ARM_MATH_MVEI) */
153
154 /**
155 @} end of mean group
156 */
157