onednn-src 0.1.13

Source of oneAPI Deep Neural Network Library (oneDNN)
Documentation 
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

/*******************************************************************************
 * Copyright 2025 Intel Corporation
 *
 * Licensed under the Apache License, Version 2.0 (the "License");
 * you may not use this file except in compliance with the License.
 * You may obtain a copy of the License at
 *
 *     http://www.apache.org/licenses/LICENSE-2.0
 *
 * 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.
 *******************************************************************************/

#ifndef GRAPH_BACKEND_DNNL_EXECUTABLES_LAYER_NORM_HPP
#define GRAPH_BACKEND_DNNL_EXECUTABLES_LAYER_NORM_HPP

#include "graph/backend/dnnl/executables/base.hpp"

namespace dnnl {
namespace impl {
namespace graph {
namespace dnnl_impl {

struct layernorm_executable_t : public op_executable_t {
    DECLARE_DESC_CLASS_AND_CREATOR(
            dnnl::layer_normalization_forward::primitive_desc);
    DECLARE_ARG_INDICES_GETTER;

    layernorm_executable_t(std::shared_ptr<op_t> &op,
            const dnnl::engine &p_engine, pd_cache_t &pd_cache,
            const fpmath_t &fpmath, bool use_block_layout) {
        auto desc
                = create_desc(op, p_engine, pd_cache, fpmath, use_block_layout);
        prim_ = dnnl::layer_normalization_forward(desc);
    }

    void execute(const stream &stream,
            const std::unordered_map<int, memory> &args) const override {
        prim_.execute(stream, args);
    }

#ifdef DNNL_WITH_SYCL
    std::optional<::sycl::event> execute_sycl(const stream &stream,
            const std::unordered_map<int, memory> &args,
            const std::vector<::sycl::event> &deps) const override {
        auto e = dnnl::sycl_interop::execute(prim_, stream, args, deps);
        if (stream.get_engine().get_kind() == engine::kind::cpu) e.wait();
        return e;
    }
#endif

#if DNNL_GPU_RUNTIME == DNNL_RUNTIME_OCL
    cl_event execute_ocl(const stream &stream,
            const std::unordered_map<int, memory> &args,
            const std::vector<cl_event> &deps) const override {
        auto e = dnnl::ocl_interop::execute(prim_, stream, args, deps);
        return e;
    }
#endif

private:
    dnnl::layer_normalization_forward prim_;
};

struct layernorm_bwd_executable_t : public op_executable_t {
    DECLARE_DESC_CLASS_AND_CREATOR(
            dnnl::layer_normalization_backward::primitive_desc);
    DECLARE_ARG_INDICES_GETTER;

    layernorm_bwd_executable_t(std::shared_ptr<op_t> &op,
            const dnnl::engine &p_engine, pd_cache_t &pd_cache,
            const fpmath_t &fpmath, bool use_block_layout) {
        auto desc
                = create_desc(op, p_engine, pd_cache, fpmath, use_block_layout);
        prim_ = dnnl::layer_normalization_backward(desc);
    }

    void execute(const stream &stream,
            const std::unordered_map<int, memory> &args) const override {
        prim_.execute(stream, args);
    }

#ifdef DNNL_WITH_SYCL
    std::optional<::sycl::event> execute_sycl(const stream &stream,
            const std::unordered_map<int, memory> &args,
            const std::vector<::sycl::event> &deps) const override {
        auto e = dnnl::sycl_interop::execute(prim_, stream, args, deps);
        if (stream.get_engine().get_kind() == engine::kind::cpu) e.wait();
        return e;
    }
#endif

#if DNNL_GPU_RUNTIME == DNNL_RUNTIME_OCL
    cl_event execute_ocl(const stream &stream,
            const std::unordered_map<int, memory> &args,
            const std::vector<cl_event> &deps) const override {
        auto e = dnnl::ocl_interop::execute(prim_, stream, args, deps);
        return e;
    }
#endif

private:
    dnnl::layer_normalization_backward prim_;
};

} // namespace dnnl_impl
} // namespace graph
} // namespace impl
} // namespace dnnl

#endif // GRAPH_BACKEND_DNNL_EXECUTABLES_LAYER_NORM_HPP