1 /* ----------------------------------------------------------------------
2 * Project: CMSIS DSP Library
3 * Title: arm_add_q31.c
4 * Description: Q31 vector addition
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 BasicAdd
37 @{
38 */
39
40 /**
41 @brief Q31 vector addition.
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
53 #if defined(ARM_MATH_MVEI) && !defined(ARM_MATH_AUTOVECTORIZE)
54
55 #include "arm_helium_utils.h"
56
arm_add_q31(const q31_t * pSrcA,const q31_t * pSrcB,q31_t * pDst,uint32_t blockSize)57 void arm_add_q31(
58 const q31_t * pSrcA,
59 const q31_t * pSrcB,
60 q31_t * pDst,
61 uint32_t blockSize)
62 {
63 uint32_t blkCnt;
64 q31x4_t vecA;
65 q31x4_t vecB;
66
67 /* Compute 4 outputs at a time */
68 blkCnt = blockSize >> 2;
69 while (blkCnt > 0U)
70 {
71 /*
72 * C = A + B
73 * Add and then store the results in the destination buffer.
74 */
75 vecA = vld1q(pSrcA);
76 vecB = vld1q(pSrcB);
77 vst1q(pDst, vqaddq(vecA, vecB));
78 /*
79 * Decrement the blockSize loop counter
80 */
81 blkCnt--;
82 /*
83 * advance vector source and destination pointers
84 */
85 pSrcA += 4;
86 pSrcB += 4;
87 pDst += 4;
88 }
89 /*
90 * tail
91 */
92 blkCnt = blockSize & 3;
93 if (blkCnt > 0U)
94 {
95 mve_pred16_t p0 = vctp32q(blkCnt);
96 vecA = vld1q(pSrcA);
97 vecB = vld1q(pSrcB);
98 vstrwq_p(pDst, vqaddq(vecA, vecB), p0);
99 }
100 }
101
102 #else
arm_add_q31(const q31_t * pSrcA,const q31_t * pSrcB,q31_t * pDst,uint32_t blockSize)103 void arm_add_q31(
104 const q31_t * pSrcA,
105 const q31_t * pSrcB,
106 q31_t * pDst,
107 uint32_t blockSize)
108 {
109 uint32_t blkCnt; /* Loop counter */
110
111 #if defined (ARM_MATH_LOOPUNROLL)
112
113 /* Loop unrolling: Compute 4 outputs at a time */
114 blkCnt = blockSize >> 2U;
115
116 while (blkCnt > 0U)
117 {
118 /* C = A + B */
119
120 /* Add and store result in destination buffer. */
121 *pDst++ = __QADD(*pSrcA++, *pSrcB++);
122
123 *pDst++ = __QADD(*pSrcA++, *pSrcB++);
124
125 *pDst++ = __QADD(*pSrcA++, *pSrcB++);
126
127 *pDst++ = __QADD(*pSrcA++, *pSrcB++);
128
129 /* Decrement loop counter */
130 blkCnt--;
131 }
132
133 /* Loop unrolling: Compute remaining outputs */
134 blkCnt = blockSize % 0x4U;
135
136 #else
137
138 /* Initialize blkCnt with number of samples */
139 blkCnt = blockSize;
140
141 #endif /* #if defined (ARM_MATH_LOOPUNROLL) */
142
143 while (blkCnt > 0U)
144 {
145 /* C = A + B */
146
147 /* Add and store result in destination buffer. */
148 *pDst++ = __QADD(*pSrcA++, *pSrcB++);
149
150 /* Decrement loop counter */
151 blkCnt--;
152 }
153
154 }
155
156 #endif /* defined(ARM_MATH_MVEI) */
157 /**
158 @} end of BasicAdd group
159 */
160
