1 /*
2 * SPDX-FileCopyrightText: Copyright 2010-2023 Arm Limited and/or its affiliates <open-source-office@arm.com>
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_relu_q7.c
22 * Description: Q7 version of ReLU
23 *
24 * $Date: 31 January 2023
25 * $Revision: V.1.2.1
26 *
27 * Target : Arm(R) M-Profile Architecture
28 *
29 * -------------------------------------------------------------------- */
30
31 #include "arm_nnfunctions.h"
32 #include "arm_nnsupportfunctions.h"
33
34 /**
35 * @ingroup Public
36 */
37
38 /**
39 * @addtogroup Acti
40 * @{
41 */
42
43 /*
44 * Q7 ReLu function
45 *
46 * Refer header file for details.
47 *
48 */
49
arm_relu_q7(int8_t * data,uint16_t size)50 void arm_relu_q7(int8_t *data, uint16_t size)
51 {
52
53 #if defined(ARM_MATH_DSP) && !defined(ARM_MATH_MVEI)
54 /* Run the following code for M cores with DSP extension */
55
56 uint16_t i = size >> 2;
57 int8_t *input = data;
58 int8_t *output = data;
59 int32_t in;
60 int32_t buf;
61 int32_t mask;
62
63 while (i)
64 {
65 in = arm_nn_read_s8x4_ia((const int8_t **)&input);
66
67 /* extract the first bit */
68 buf = (int32_t)ROR((uint32_t)in & 0x80808080, 7);
69
70 /* if MSB=1, mask will be 0xFF, 0x0 otherwise */
71 mask = QSUB8(0x00000000, buf);
72
73 arm_nn_write_s8x4_ia(&output, in & (~mask));
74
75 i--;
76 }
77
78 i = size & 0x3;
79 while (i)
80 {
81 if (*input < 0)
82 {
83 *input = 0;
84 }
85 input++;
86 i--;
87 }
88
89 #else
90 /* Run the following code as reference implementation for cores without DSP extension */
91
92 uint16_t i;
93
94 for (i = 0; i < size; i++)
95 {
96 if (data[i] < 0)
97 data[i] = 0;
98 }
99
100 #endif
101 }
102
103 /**
104 * @} end of Acti group
105 */
106
