1 /* ----------------------------------------------------------------------
2 * Project: CMSIS DSP Library
3 * Title: arm_mult_q31.c
4 * Description: Q31 vector multiplication
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/basic_math_functions.h"
30
31 /**
32 @ingroup groupMath
33 */
34
35 /**
36 @addtogroup BasicMult
37 @{
38 */
39
40 /**
41 @brief Q31 vector multiplication.
42 @param[in] pSrcA points to the first input vector.
43 @param[in] pSrcB points to the second input vector.
44 @param[out] pDst points to the output vector.
45 @param[in] blockSize number of samples in each vector.
46 @return none
47
48 @par Scaling and Overflow Behavior
49 The function uses saturating arithmetic.
50 Results outside of the allowable Q31 range[0x80000000 0x7FFFFFFF] are saturated.
51 */
52 #if defined(ARM_MATH_MVEI) && !defined(ARM_MATH_AUTOVECTORIZE)
53
54 #include "arm_helium_utils.h"
55
arm_mult_q31(const q31_t * pSrcA,const q31_t * pSrcB,q31_t * pDst,uint32_t blockSize)56 void arm_mult_q31(
57 const q31_t * pSrcA,
58 const q31_t * pSrcB,
59 q31_t * pDst,
60 uint32_t blockSize)
61 {
62 uint32_t blkCnt; /* loop counters */
63 q31x4_t vecA, vecB;
64
65 /* Compute 4 outputs at a time */
66 blkCnt = blockSize >> 2;
67 while (blkCnt > 0U)
68 {
69 /*
70 * C = A * B
71 * Multiply the inputs and then store the results in the destination buffer.
72 */
73 vecA = vld1q(pSrcA);
74 vecB = vld1q(pSrcB);
75 vst1q(pDst, vqdmulhq(vecA, vecB));
76 /*
77 * Decrement the blockSize loop counter
78 */
79 blkCnt--;
80 /*
81 * advance vector source and destination pointers
82 */
83 pSrcA += 4;
84 pSrcB += 4;
85 pDst += 4;
86 }
87 /*
88 * tail
89 */
90 blkCnt = blockSize & 3;
91 if (blkCnt > 0U)
92 {
93 mve_pred16_t p0 = vctp32q(blkCnt);
94 vecA = vld1q(pSrcA);
95 vecB = vld1q(pSrcB);
96 vstrwq_p(pDst, vqdmulhq(vecA, vecB), p0);
97 }
98 }
99
100 #else
arm_mult_q31(const q31_t * pSrcA,const q31_t * pSrcB,q31_t * pDst,uint32_t blockSize)101 void arm_mult_q31(
102 const q31_t * pSrcA,
103 const q31_t * pSrcB,
104 q31_t * pDst,
105 uint32_t blockSize)
106 {
107 uint32_t blkCnt; /* Loop counter */
108 q31_t out; /* Temporary output variable */
109
110 #if defined (ARM_MATH_LOOPUNROLL)
111
112 /* Loop unrolling: Compute 4 outputs at a time */
113 blkCnt = blockSize >> 2U;
114
115 while (blkCnt > 0U)
116 {
117 /* C = A * B */
118
119 /* Multiply inputs and store result in destination buffer. */
120 out = ((q63_t) *pSrcA++ * *pSrcB++) >> 32;
121 out = __SSAT(out, 31);
122 *pDst++ = out << 1U;
123
124 out = ((q63_t) *pSrcA++ * *pSrcB++) >> 32;
125 out = __SSAT(out, 31);
126 *pDst++ = out << 1U;
127
128 out = ((q63_t) *pSrcA++ * *pSrcB++) >> 32;
129 out = __SSAT(out, 31);
130 *pDst++ = out << 1U;
131
132 out = ((q63_t) *pSrcA++ * *pSrcB++) >> 32;
133 out = __SSAT(out, 31);
134 *pDst++ = out << 1U;
135
136 /* Decrement loop counter */
137 blkCnt--;
138 }
139
140 /* Loop unrolling: Compute remaining outputs */
141 blkCnt = blockSize % 0x4U;
142
143 #else
144
145 /* Initialize blkCnt with number of samples */
146 blkCnt = blockSize;
147
148 #endif /* #if defined (ARM_MATH_LOOPUNROLL) */
149
150 while (blkCnt > 0U)
151 {
152 /* C = A * B */
153
154 /* Multiply inputs and store result in destination buffer. */
155 out = ((q63_t) *pSrcA++ * *pSrcB++) >> 32;
156 out = __SSAT(out, 31);
157 *pDst++ = out << 1U;
158
159 /* Decrement loop counter */
160 blkCnt--;
161 }
162
163 }
164 #endif /* defined(ARM_MATH_MVEI) */
165
166 /**
167 @} end of BasicMult group
168 */
169
