1 /* ----------------------------------------------------------------------
2 * Project: CMSIS DSP Library
3 * Title: arm_cmplx_mag_squared_q31.c
4 * Description: Q31 complex magnitude squared
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/complex_math_functions.h"
30
31 /**
32 @ingroup groupCmplxMath
33 */
34
35 /**
36 @addtogroup cmplx_mag_squared
37 @{
38 */
39
40 /**
41 @brief Q31 complex magnitude squared.
42 @param[in] pSrc points to input vector
43 @param[out] pDst points to output vector
44 @param[in] numSamples number of samples in each vector
45
46 @par Scaling and Overflow Behavior
47 The function implements 1.31 by 1.31 multiplications and finally output is converted into 3.29 format.
48 Input down scaling is not required.
49 */
50
51 #if defined(ARM_MATH_MVEI) && !defined(ARM_MATH_AUTOVECTORIZE)
52
arm_cmplx_mag_squared_q31(const q31_t * pSrc,q31_t * pDst,uint32_t numSamples)53 void arm_cmplx_mag_squared_q31(
54 const q31_t * pSrc,
55 q31_t * pDst,
56 uint32_t numSamples)
57 {
58 int32_t blockSize = numSamples; /* loop counters */
59 uint32_t blkCnt; /* loop counters */
60 q31x4x2_t vecSrc;
61 q31x4_t vReal, vImag;
62 q31x4_t vMagSq;
63 q31_t real, imag; /* Temporary input variables */
64 q31_t acc0, acc1; /* Accumulators */
65
66 /* Compute 4 complex samples at a time */
67 blkCnt = blockSize >> 2;
68 while (blkCnt > 0U)
69 {
70 vecSrc = vld2q(pSrc);
71 vReal = vmulhq(vecSrc.val[0], vecSrc.val[0]);
72 vImag = vmulhq(vecSrc.val[1], vecSrc.val[1]);
73 vMagSq = vqaddq(vReal, vImag);
74 vMagSq = vshrq(vMagSq, 1);
75
76 vst1q(pDst, vMagSq);
77
78 pSrc += 8;
79 pDst += 4;
80 /*
81 * Decrement the blkCnt loop counter
82 * Advance vector source and destination pointers
83 */
84 blkCnt --;
85 }
86
87 /* Tail */
88 blkCnt = blockSize & 3;
89 while (blkCnt > 0U)
90 {
91 /* C[0] = (A[0] * A[0] + A[1] * A[1]) */
92
93 real = *pSrc++;
94 imag = *pSrc++;
95 acc0 = (q31_t) (((q63_t) real * real) >> 33);
96 acc1 = (q31_t) (((q63_t) imag * imag) >> 33);
97
98 /* store result in 3.29 format in destination buffer. */
99 *pDst++ = acc0 + acc1;
100
101 /* Decrement loop counter */
102 blkCnt--;
103 }
104 }
105
106 #else
arm_cmplx_mag_squared_q31(const q31_t * pSrc,q31_t * pDst,uint32_t numSamples)107 void arm_cmplx_mag_squared_q31(
108 const q31_t * pSrc,
109 q31_t * pDst,
110 uint32_t numSamples)
111 {
112 uint32_t blkCnt; /* Loop counter */
113 q31_t real, imag; /* Temporary input variables */
114 q31_t acc0, acc1; /* Accumulators */
115
116 #if defined (ARM_MATH_LOOPUNROLL)
117
118 /* Loop unrolling: Compute 4 outputs at a time */
119 blkCnt = numSamples >> 2U;
120
121 while (blkCnt > 0U)
122 {
123 /* C[0] = (A[0] * A[0] + A[1] * A[1]) */
124
125 real = *pSrc++;
126 imag = *pSrc++;
127 acc0 = (q31_t) (((q63_t) real * real) >> 33);
128 acc1 = (q31_t) (((q63_t) imag * imag) >> 33);
129 /* store the result in 3.29 format in the destination buffer. */
130 *pDst++ = acc0 + acc1;
131
132 real = *pSrc++;
133 imag = *pSrc++;
134 acc0 = (q31_t) (((q63_t) real * real) >> 33);
135 acc1 = (q31_t) (((q63_t) imag * imag) >> 33);
136 *pDst++ = acc0 + acc1;
137
138 real = *pSrc++;
139 imag = *pSrc++;
140 acc0 = (q31_t) (((q63_t) real * real) >> 33);
141 acc1 = (q31_t) (((q63_t) imag * imag) >> 33);
142 *pDst++ = acc0 + acc1;
143
144 real = *pSrc++;
145 imag = *pSrc++;
146 acc0 = (q31_t) (((q63_t) real * real) >> 33);
147 acc1 = (q31_t) (((q63_t) imag * imag) >> 33);
148 *pDst++ = acc0 + acc1;
149
150 /* Decrement loop counter */
151 blkCnt--;
152 }
153
154 /* Loop unrolling: Compute remaining outputs */
155 blkCnt = numSamples % 0x4U;
156
157 #else
158
159 /* Initialize blkCnt with number of samples */
160 blkCnt = numSamples;
161
162 #endif /* #if defined (ARM_MATH_LOOPUNROLL) */
163
164 while (blkCnt > 0U)
165 {
166 /* C[0] = (A[0] * A[0] + A[1] * A[1]) */
167
168 real = *pSrc++;
169 imag = *pSrc++;
170 acc0 = (q31_t) (((q63_t) real * real) >> 33);
171 acc1 = (q31_t) (((q63_t) imag * imag) >> 33);
172
173 /* store result in 3.29 format in destination buffer. */
174 *pDst++ = acc0 + acc1;
175
176 /* Decrement loop counter */
177 blkCnt--;
178 }
179
180 }
181
182 #endif /* defined(ARM_MATH_MVEI) */
183
184 /**
185 @} end of cmplx_mag_squared group
186 */
187
