1 /*
2 * Copyright (C) 2010-2018 Arm Limited or its affiliates. All rights reserved.
3 *
4 * SPDX-License-Identifier: Apache-2.0
5 *
6 * Licensed under the Apache License, Version 2.0 (the License); you may
7 * not use this file except in compliance with the License.
8 * You may obtain a copy of the License at
9 *
10 * www.apache.org/licenses/LICENSE-2.0
11 *
12 * Unless required by applicable law or agreed to in writing, software
13 * distributed under the License is distributed on an AS IS BASIS, WITHOUT
14 * WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
15 * See the License for the specific language governing permissions and
16 * limitations under the License.
17 */
18
19 /* ----------------------------------------------------------------------
20 * Project: CMSIS NN Library
21 * Title: arm_nn_add_q7.c
22 * Description: Non saturating addition of elements of a q7 vector.
23 *
24 * $Date: 09. October 2020
25 * $Revision: V.1.0.1
26 *
27 * Target Processor: Cortex-M cores
28 *
29 * -------------------------------------------------------------------- */
30
31 #include "arm_nn_tables.h"
32 #include "arm_nnsupportfunctions.h"
33
34 /**
35 * @ingroup groupSupport
36 */
37
38 /**
39 * @addtogroup NNBasicMath
40 * @{
41 */
42
arm_nn_add_q7(const q7_t * input,q31_t * output,uint32_t block_size)43 void arm_nn_add_q7(const q7_t *input, q31_t *output, uint32_t block_size)
44 {
45 uint32_t block_count;
46 q31_t result = 0;
47 #if defined(ARM_MATH_DSP)
48 /* Loop unrolling: Compute 4 outputs at a time */
49 block_count = block_size >> 2U;
50
51 while (block_count > 0U)
52 {
53 const int32_t mult_q15x2 = (1UL << 16) | 1UL;
54 q31_t in_q7x4 = arm_nn_read_q7x4_ia(&input);
55 q31_t temp_q15x2 = __SXTAB16(__SXTB16(in_q7x4), __ROR((uint32_t)in_q7x4, 8));
56
57 result = __SMLAD(temp_q15x2, mult_q15x2, result);
58
59 /* Decrement loop counter */
60 block_count--;
61 }
62
63 /* Loop unrolling: Compute remaining outputs */
64 block_count = block_size & 0x3;
65 #else
66 block_count = block_size;
67 #endif
68 while (block_count > 0U)
69 {
70 /* Add and store result in destination buffer. */
71 result += *input++;
72
73 /* Decrement loop counter */
74 block_count--;
75 }
76
77 *output = result;
78 }
79
80 /**
81 * @} end of NNBasicMath group
82 */
