1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
|
/* ----------------------------------------------------------------------
* Project: CMSIS DSP Library
* Title: arm_cmplx_mult_real_q15.c
* Description: Q15 complex by real multiplication
*
* $Date: 23 April 2021
* $Revision: V1.9.0
*
* Target Processor: Cortex-M and Cortex-A cores
* -------------------------------------------------------------------- */
/*
* Copyright (C) 2010-2021 ARM Limited or its affiliates. All rights reserved.
*
* SPDX-License-Identifier: Apache-2.0
*
* Licensed under the Apache License, Version 2.0 (the License); you may
* not use this file except in compliance with the License.
* You may obtain a copy of the License at
*
* www.apache.org/licenses/LICENSE-2.0
*
* Unless required by applicable law or agreed to in writing, software
* distributed under the License is distributed on an AS IS BASIS, WITHOUT
* WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
* See the License for the specific language governing permissions and
* limitations under the License.
*/
#include "dsp/complex_math_functions.h"
/**
@ingroup groupCmplxMath
*/
/**
@addtogroup CmplxByRealMult
@{
*/
/**
@brief Q15 complex-by-real multiplication.
@param[in] pSrcCmplx points to complex input vector
@param[in] pSrcReal points to real input vector
@param[out] pCmplxDst points to complex output vector
@param[in] numSamples number of samples in each vector
@return none
@par Scaling and Overflow Behavior
The function uses saturating arithmetic.
Results outside of the allowable Q15 range [0x8000 0x7FFF] are saturated.
*/
#if defined(ARM_MATH_MVEI) && !defined(ARM_MATH_AUTOVECTORIZE)
void arm_cmplx_mult_real_q15(
const q15_t * pSrcCmplx,
const q15_t * pSrcReal,
q15_t * pCmplxDst,
uint32_t numSamples)
{
static const uint16_t stride_cmplx_x_real_16[8] = {
0, 0, 1, 1, 2, 2, 3, 3
};
q15x8_t rVec;
q15x8_t cmplxVec;
q15x8_t dstVec;
uint16x8_t strideVec;
uint32_t blockSizeC = numSamples * CMPLX_DIM; /* loop counters */
uint32_t blkCnt;
q15_t in;
/*
* stride vector for pairs of real generation
*/
strideVec = vld1q(stride_cmplx_x_real_16);
blkCnt = blockSizeC >> 3;
while (blkCnt > 0U)
{
cmplxVec = vld1q(pSrcCmplx);
rVec = vldrhq_gather_shifted_offset_s16(pSrcReal, strideVec);
dstVec = vqdmulhq(cmplxVec, rVec);
vst1q(pCmplxDst, dstVec);
pSrcReal += 4;
pSrcCmplx += 8;
pCmplxDst += 8;
blkCnt --;
}
/* Tail */
blkCnt = (blockSizeC & 7) >> 1;
while (blkCnt > 0U)
{
/* C[2 * i ] = A[2 * i ] * B[i]. */
/* C[2 * i + 1] = A[2 * i + 1] * B[i]. */
in = *pSrcReal++;
/* store the result in the destination buffer. */
*pCmplxDst++ = (q15_t) __SSAT((((q31_t) *pSrcCmplx++ * in) >> 15), 16);
*pCmplxDst++ = (q15_t) __SSAT((((q31_t) *pSrcCmplx++ * in) >> 15), 16);
/* Decrement loop counter */
blkCnt--;
}
}
#else
void arm_cmplx_mult_real_q15(
const q15_t * pSrcCmplx,
const q15_t * pSrcReal,
q15_t * pCmplxDst,
uint32_t numSamples)
{
uint32_t blkCnt; /* Loop counter */
q15_t in; /* Temporary variable */
#if defined (ARM_MATH_LOOPUNROLL)
#if defined (ARM_MATH_DSP)
q31_t inA1, inA2; /* Temporary variables to hold input data */
q31_t inB1; /* Temporary variables to hold input data */
q15_t out1, out2, out3, out4; /* Temporary variables to hold output data */
q31_t mul1, mul2, mul3, mul4; /* Temporary variables to hold intermediate data */
#endif
/* Loop unrolling: Compute 4 outputs at a time */
blkCnt = numSamples >> 2U;
while (blkCnt > 0U)
{
/* C[2 * i ] = A[2 * i ] * B[i]. */
/* C[2 * i + 1] = A[2 * i + 1] * B[i]. */
#if defined (ARM_MATH_DSP)
/* read 2 complex numbers both real and imaginary from complex input buffer */
inA1 = read_q15x2_ia (&pSrcCmplx);
inA2 = read_q15x2_ia (&pSrcCmplx);
/* read 2 real values at a time from real input buffer */
inB1 = read_q15x2_ia (&pSrcReal);
/* multiply complex number with real numbers */
#ifndef ARM_MATH_BIG_ENDIAN
mul1 = (q31_t) ((q15_t) (inA1) * (q15_t) (inB1));
mul2 = (q31_t) ((q15_t) (inA1 >> 16) * (q15_t) (inB1));
mul3 = (q31_t) ((q15_t) (inA2) * (q15_t) (inB1 >> 16));
mul4 = (q31_t) ((q15_t) (inA2 >> 16) * (q15_t) (inB1 >> 16));
#else
mul2 = (q31_t) ((q15_t) (inA1 >> 16) * (q15_t) (inB1 >> 16));
mul1 = (q31_t) ((q15_t) inA1 * (q15_t) (inB1 >> 16));
mul4 = (q31_t) ((q15_t) (inA2 >> 16) * (q15_t) inB1);
mul3 = (q31_t) ((q15_t) inA2 * (q15_t) inB1);
#endif /* #ifndef ARM_MATH_BIG_ENDIAN */
/* saturate the result */
out1 = (q15_t) __SSAT(mul1 >> 15U, 16);
out2 = (q15_t) __SSAT(mul2 >> 15U, 16);
out3 = (q15_t) __SSAT(mul3 >> 15U, 16);
out4 = (q15_t) __SSAT(mul4 >> 15U, 16);
/* pack real and imaginary outputs and store them to destination */
write_q15x2_ia (&pCmplxDst, __PKHBT(out1, out2, 16));
write_q15x2_ia (&pCmplxDst, __PKHBT(out3, out4, 16));
inA1 = read_q15x2_ia (&pSrcCmplx);
inA2 = read_q15x2_ia (&pSrcCmplx);
inB1 = read_q15x2_ia (&pSrcReal);
#ifndef ARM_MATH_BIG_ENDIAN
mul1 = (q31_t) ((q15_t) (inA1) * (q15_t) (inB1));
mul2 = (q31_t) ((q15_t) (inA1 >> 16) * (q15_t) (inB1));
mul3 = (q31_t) ((q15_t) (inA2) * (q15_t) (inB1 >> 16));
mul4 = (q31_t) ((q15_t) (inA2 >> 16) * (q15_t) (inB1 >> 16));
#else
mul2 = (q31_t) ((q15_t) (inA1 >> 16) * (q15_t) (inB1 >> 16));
mul1 = (q31_t) ((q15_t) inA1 * (q15_t) (inB1 >> 16));
mul4 = (q31_t) ((q15_t) (inA2 >> 16) * (q15_t) inB1);
mul3 = (q31_t) ((q15_t) inA2 * (q15_t) inB1);
#endif /* #ifndef ARM_MATH_BIG_ENDIAN */
out1 = (q15_t) __SSAT(mul1 >> 15U, 16);
out2 = (q15_t) __SSAT(mul2 >> 15U, 16);
out3 = (q15_t) __SSAT(mul3 >> 15U, 16);
out4 = (q15_t) __SSAT(mul4 >> 15U, 16);
write_q15x2_ia (&pCmplxDst, __PKHBT(out1, out2, 16));
write_q15x2_ia (&pCmplxDst, __PKHBT(out3, out4, 16));
#else
in = *pSrcReal++;
*pCmplxDst++ = (q15_t) __SSAT((((q31_t) *pSrcCmplx++ * in) >> 15), 16);
*pCmplxDst++ = (q15_t) __SSAT((((q31_t) *pSrcCmplx++ * in) >> 15), 16);
in = *pSrcReal++;
*pCmplxDst++ = (q15_t) __SSAT((((q31_t) *pSrcCmplx++ * in) >> 15), 16);
*pCmplxDst++ = (q15_t) __SSAT((((q31_t) *pSrcCmplx++ * in) >> 15), 16);
in = *pSrcReal++;
*pCmplxDst++ = (q15_t) __SSAT((((q31_t) *pSrcCmplx++ * in) >> 15), 16);
*pCmplxDst++ = (q15_t) __SSAT((((q31_t) *pSrcCmplx++ * in) >> 15), 16);
in = *pSrcReal++;
*pCmplxDst++ = (q15_t) __SSAT((((q31_t) *pSrcCmplx++ * in) >> 15), 16);
*pCmplxDst++ = (q15_t) __SSAT((((q31_t) *pSrcCmplx++ * in) >> 15), 16);
#endif
/* Decrement loop counter */
blkCnt--;
}
/* Loop unrolling: Compute remaining outputs */
blkCnt = numSamples % 0x4U;
#else
/* Initialize blkCnt with number of samples */
blkCnt = numSamples;
#endif /* #if defined (ARM_MATH_LOOPUNROLL) */
while (blkCnt > 0U)
{
/* C[2 * i ] = A[2 * i ] * B[i]. */
/* C[2 * i + 1] = A[2 * i + 1] * B[i]. */
in = *pSrcReal++;
/* store the result in the destination buffer. */
*pCmplxDst++ = (q15_t) __SSAT((((q31_t) *pSrcCmplx++ * in) >> 15), 16);
*pCmplxDst++ = (q15_t) __SSAT((((q31_t) *pSrcCmplx++ * in) >> 15), 16);
/* Decrement loop counter */
blkCnt--;
}
}
#endif /* defined(ARM_MATH_MVEI) */
/**
@} end of CmplxByRealMult group
*/
|
