File size: 3,381 Bytes
be903e2
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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
// Tencent is pleased to support the open source community by making ncnn available.
//
// Copyright (C) 2019 THL A29 Limited, a Tencent company. All rights reserved.
//
// Licensed under the BSD 3-Clause License (the "License"); you may not use this file except
// in compliance with the License. You may obtain a copy of the License at
//
// https://opensource.org/licenses/BSD-3-Clause
//
// 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 "cast_x86_avx.h"

#if __SSE2__
#include <emmintrin.h>
#if __AVX__
#include <immintrin.h>
#endif // __AVX__
#endif // __SSE2__
#include "x86_usability.h"

#include "cpu.h"

namespace ncnn {

#include "cast_fp16.h"
#include "cast_bf16.h"

Cast_x86_avx::Cast_x86_avx()
{
    support_packing = true;
}

int Cast_x86_avx::forward(const Mat& bottom_blob, Mat& top_blob, const Option& opt) const
{
    if (type_from == type_to)
    {
        top_blob = bottom_blob;
        return 0;
    }

    int w = bottom_blob.w;
    int h = bottom_blob.h;
    int d = bottom_blob.d;
    int channels = bottom_blob.c;
    int dims = bottom_blob.dims;
    size_t elemsize = bottom_blob.elemsize;
    int elempack = bottom_blob.elempack;

    size_t out_elemsize = elemsize;
    if (type_to == 1)
    {
        if (type_from == 3)
        {
            Cast::forward(bottom_blob, top_blob, opt);
        }

        // float32
        out_elemsize = 4 * elempack;
    }
    else if (type_to == 2)
    {
        // float16
        out_elemsize = 2 * elempack;
    }
    else if (type_to == 3)
    {
        // int8
        out_elemsize = elempack;
    }
    else if (type_to == 4)
    {
        // bfloat16
        out_elemsize = 2 * elempack;
    }

    if (dims == 1)
    {
        top_blob.create(w, out_elemsize, elempack, opt.blob_allocator);
    }
    else if (dims == 2)
    {
        top_blob.create(w, h, out_elemsize, elempack, opt.blob_allocator);
    }
    else if (dims == 3)
    {
        top_blob.create(w, h, channels, out_elemsize, elempack, opt.blob_allocator);
    }
    else if (dims == 4)
    {
        top_blob.create(w, h, d, channels, out_elemsize, elempack, opt.blob_allocator);
    }
    if (top_blob.empty())
        return -100;

    int size = w * h * d * elempack;

    if (type_from == 1 && type_to == 2)
    {
        cast_fp32_to_fp16_sse(bottom_blob, top_blob, opt);
    }

    if (type_from == 2 && type_to == 1)
    {
        cast_fp16_to_fp32_sse(bottom_blob, top_blob, opt);
    }

    if (type_from == 3 && type_to == 1)
    {
        #pragma omp parallel for num_threads(opt.num_threads)
        for (int q = 0; q < channels; q++)
        {
            const signed char* ptr = bottom_blob.channel(q);
            float* outptr = top_blob.channel(q);

            for (int i = 0; i < size; i++)
            {
                outptr[i] = (float)ptr[i];
            }
        }
    }

    if (type_from == 1 && type_to == 4)
    {
        cast_fp32_to_bf16_sse(bottom_blob, top_blob, opt);
    }

    if (type_from == 4 && type_to == 1)
    {
        cast_bf16_to_fp32_sse(bottom_blob, top_blob, opt);
    }

    return 0;
}

} // namespace ncnn