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
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
141
142
143
144
145
146
147
148
149
150

/*******************************************************************************
* Copyright 2019 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.
*******************************************************************************/

#include "gpu/intel/rnn/config.hpp"
#include "gpu/intel/rnn/grid.hpp"
// Common for RNN and LSTM cell execution
#include "gpu/intel/rnn/simple_cell_fusion.hpp"

namespace dnnl {
namespace impl {
namespace gpu {
namespace intel {
namespace rnn {

using namespace utils;

template <prop_kind_t aprop>
cell_execution_sig((simple_common_t<aprop>::cell_execution)) {
    const conf_t &conf = this->pd()->conf;
    const ocl_conf_t &ocl_conf = this->pd()->ocl_conf;
    const offsets_t &offsets = this->pd()->off;

    const bool use_cell = ocl_conf.cell_comp.is_enabled;

    strides_t<4> user_layer_strides {[&]() {
        auto s = user_data.src_layer_strides(dir);
        return strides_t<4> {0, 0, s[0], s[1]};
    }()};

    auto cell_layer = !conf.copy_src_layer && lay == 0
            ? user_data.src_layer(dir, iter)
            : workspace.states(lay - 1, dir, iter);
    auto &cell_layer_strides = !conf.copy_src_layer && lay == 0
            ? user_layer_strides
            : workspace.states_strides();
    auto cell_iter = workspace.states(lay, dir, iter - 1);
    auto &cell_iter_strides = workspace.states_strides();
    auto scratch_gates = scratch.gates(iter);
    strides_t<2> scratch_gates_strides
            = {scratch.calc_off_gates(1), conf.scratch_gates_ld};

    auto wei_layer = user_data.wei_layer(lay, dir);
    auto wei_iter = user_data.wei_iter(lay, dir);
    // passing empty memory for scratch cell since its not used in lstm and rnn
    auto &scratch_cell = memory_storage_t::empty_storage();

    if ((aprop == prop_kind::forward) || conf.recompute_gates) {
        if (!conf.merge_gemm_layer && !conf.cell_fusion.gemm_layer) {
            auto gemm_cell_layer_fwd = !conf.copy_src_layer && lay == 0
                    ? gemm_layer_fwd_src
                    : gemm_layer_fwd;
            CHECK(gemm_primitive(engine, ctx, wei_layer, cell_layer,
                    scratch_gates, gemm_cell_layer_fwd));
        }

        if (!conf.cell_fusion.gemm_iter)
            CHECK(gemm_primitive(engine, ctx, wei_iter, cell_iter,
                    scratch_gates, gemm_iter_fwd));
    }

    if (aprop == prop_kind::forward) {
        if (!use_cell) {
            CHECK((this->*elemwise_common)(ctx, dir, lay, iter, conf.dhc,
                    conf.mb, 1, user_data, workspace, scratch_gates, {}, {}, {},
                    {}, {}, {}, 0, scales, tm_scales, diff_bias));
        } else {
            CHECK(compute_cell_fwd(ctx, kernels_[kernel_id::cell_fwd], lay, dir,
                    iter, workspace, user_data, wei_layer, wei_iter, cell_layer,
                    cell_layer_strides, cell_iter, cell_iter_strides,
                    scratch_gates, scratch_gates_strides, scratch_cell,
                    pd()->desc()->alpha, tm_scales, conf, ocl_conf, offsets));
        }

    } else { // backward
        auto diff_states_iter = scratch.diff_states(lay, dir, 0, iter + 1);
        auto diff_states_iter_s1 = conf.n_states == 2
                ? scratch.diff_states(lay, dir, 1, iter + 1)
                : sub_buffer_t();
        auto diff_states_layer
                = !conf.copy_diff_dst_layer && lay + 1 == conf.n_layer
                ? user_data.diff_dst_layer(dir, iter)
                : scratch.diff_states(lay + 1, dir, conf.n_states, iter);
        auto diff_states_layer_ld
                = !conf.copy_diff_dst_layer && lay + 1 == conf.n_layer
                ? offsets.diff_dst_layer[1]
                : conf.scratch_diff_states_ld;

        auto diff_states = scratch.diff_states(lay, dir, 0, iter);
        auto diff_states_s1 = conf.n_states == 2
                ? scratch.diff_states(lay, dir, 1, iter)
                : sub_buffer_t();
        auto diff_states1 = !conf.copy_diff_src_layer && lay == 0
                ? user_data.diff_src_layer(dir, iter)
                : scratch.diff_states(lay, dir, conf.n_states, iter);
        auto diff_gates = scratch.diff_gates(iter);

        CHECK((this->*elemwise_common)(ctx, dir, lay, iter, conf.dhc, conf.mb,
                ocl_conf.elemwise_bwd_batch_block, user_data, workspace,
                scratch_gates, diff_gates, diff_states, diff_states_s1,
                diff_states_iter, diff_states_iter_s1, diff_states_layer,
                diff_states_layer_ld, scales, tm_scales, diff_bias));

        CHECK(gemm_primitive(
                engine, ctx, wei_iter, diff_gates, diff_states, gemm_iter_bwd));

        if (!conf.merge_gemm_layer) {

            auto gemm_layer_cell_bwd = !conf.copy_diff_src_layer && lay == 0
                    ? gemm_layer_bwd_src
                    : gemm_layer_bwd;
            CHECK(gemm_primitive(engine, ctx, wei_layer, diff_gates,
                    diff_states1, gemm_layer_cell_bwd));

            auto gemm_diff_wei_cell_layer = !conf.copy_src_layer && lay == 0
                    ? gemm_diff_wei_layer_src
                    : gemm_diff_wei_layer;

            CHECK(gemm_primitive(engine, ctx, diff_gates, cell_layer,
                    user_data.diff_wei_layer(lay, dir),
                    gemm_diff_wei_cell_layer));
        }

        if (!conf.merge_gemm_iter) {
            CHECK(gemm_primitive(engine, ctx, diff_gates, cell_iter,
                    user_data.diff_wei_iter(lay, dir), gemm_diff_wei_iter));
        }
    }
    return status::success;
}
template cell_execution_sig(simple_fwd_t::cell_execution);
template cell_execution_sig(simple_bwd_t::cell_execution);

} // namespace rnn
} // namespace intel
} // namespace gpu
} // namespace impl
} // namespace dnnl