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
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205

/*******************************************************************************
* Copyright 2020 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 <math.h>

#include "common/c_types_map.hpp"
#include "common/compiler_workarounds.hpp"
#include "common/dnnl_thread.hpp"
#include "common/nstl.hpp"

#include "cpu/simple_q10n.hpp"

#include "cpu/ref_io_helper.hpp"
#include "cpu/ref_reduction.hpp"

namespace dnnl {
namespace impl {
namespace cpu {

template <typename T>
void init_acc(T &acc, alg_kind_t alg, data_type_t src_dt) {
    using namespace alg_kind;
    using namespace nstl;

    switch (alg) {
        case reduction_max: acc = types::lowest_value<T>(src_dt); break;
        case reduction_min: acc = types::max_value<T>(src_dt); break;
        case reduction_mean:
        case reduction_sum: acc = T(0); break;
        case reduction_mul: acc = T(1); break;
        case reduction_norm_lp_max:
        case reduction_norm_lp_sum:
        case reduction_norm_lp_power_p_max:
        case reduction_norm_lp_power_p_sum: acc = T(0); break;
        default: assert(!"unknown alg");
    }
}

template <typename T>
void accumulate(T &acc, const T &src, alg_kind_t alg, float p) {
    using namespace alg_kind;

    switch (alg) {
        case reduction_max: acc = nstl::max(acc, src); break;
        case reduction_min: acc = nstl::min(acc, src); break;
        case reduction_mean:
        case reduction_sum: acc += src; break;
        case reduction_mul: acc *= src; break;
        case reduction_norm_lp_max:
        case reduction_norm_lp_sum:
        case reduction_norm_lp_power_p_max:
        case reduction_norm_lp_power_p_sum:
            acc += powf(nstl::abs(src), p);
            break;
        default: assert(!"unknown alg");
    }
}

void finalize(float &acc_f32, alg_kind_t alg, float p, float eps, dim_t n) {
    using namespace alg_kind;

    switch (alg) {
        case reduction_mean: acc_f32 /= n; break;
        case reduction_norm_lp_max:
            acc_f32 = nstl::max(acc_f32, eps);
            acc_f32 = powf(acc_f32, 1.0f / p);
            break;
        case reduction_norm_lp_sum:
            acc_f32 += eps;
            acc_f32 = powf(acc_f32, 1.0f / p);
            break;
        case reduction_norm_lp_power_p_max:
            acc_f32 = nstl::max(acc_f32, eps);
            break;
        case reduction_norm_lp_power_p_sum: acc_f32 += eps; break;
        default: break;
    }
}

template <typename T>
T ker(dim_t l_offset, alg_kind_t alg, float p, dim_t reduce_size,
        const dims_t idle_pos, const dims_t reduce_dims,
        const memory_desc_t *src_md, const void *src);

template <>
int ker(dim_t l_offset, alg_kind_t alg, float p, dim_t reduce_size,
        const dims_t idle_pos, const dims_t reduce_dims,
        const memory_desc_t *src_md, const void *src) {
    const memory_desc_wrapper src_mdw(src_md);
    const int ndims = src_mdw.ndims();
    const dim_t src_idle_off = src_mdw.off_v(idle_pos);
    dims_t reduce_pos;

    int acc {0};
    init_acc(acc, alg, src_mdw.data_type());
    for (dim_t r = 0; r < reduce_size; ++r) {
        utils::l_dims_by_l_offset(reduce_pos, r, reduce_dims, ndims);
        const dim_t src_reduce_off = src_mdw.off_v(reduce_pos);
        const dim_t src_off = src_idle_off + src_reduce_off;
        const int s = io::load_int_value(src_mdw.data_type(), src, src_off);
        accumulate(acc, s, alg, p);
    }

    return acc;
}

template <>
float ker(dim_t l_offset, alg_kind_t alg, float p, dim_t reduce_size,
        const dims_t idle_pos, const dims_t reduce_dims,
        const memory_desc_t *src_md, const void *src) {
    const memory_desc_wrapper src_mdw(src_md);
    const int ndims = src_mdw.ndims();
    const dim_t src_idle_off = src_mdw.off_v(idle_pos);
    dims_t reduce_pos;

    float acc {0};
    init_acc(acc, alg, src_mdw.data_type());
    for (dim_t r = 0; r < reduce_size; ++r) {
        utils::l_dims_by_l_offset(reduce_pos, r, reduce_dims, ndims);
        const dim_t src_reduce_off = src_mdw.off_v(reduce_pos);
        const dim_t src_off = src_idle_off + src_reduce_off;
        const float s = io::load_float_value(src_mdw.data_type(), src, src_off);
        accumulate(acc, s, alg, p);
    }

    return acc;
}

status_t ref_reduction_t::execute_ref(const exec_ctx_t &ctx) const {
    status_t status = status::success;
    auto src = CTX_IN_MEM(const void *, DNNL_ARG_SRC);
    auto dst = CTX_OUT_CLEAN_MEM(void *, DNNL_ARG_DST, status);
    CHECK(status);

    const memory_desc_wrapper src_mdw(pd()->src_md());
    const memory_desc_wrapper dst_mdw(pd()->dst_md());

    const int ndims = src_mdw.ndims();
    const auto &src_dims = src_mdw.dims();
    const auto &dst_dims = dst_mdw.dims();
    const auto acc_type = types::default_accum_data_type(
            src_mdw.data_type(), dst_mdw.data_type());

    const auto alg = pd()->desc()->alg_kind;
    const auto p = pd()->desc()->p;
    const auto eps = pd()->desc()->eps;

    dims_t reduce_dims;
    dim_t reduce_size {1}, idle_size = dst_mdw.nelems();

    for (int d = 0; d < ndims; ++d) {
        reduce_dims[d] = dim_t {1};
        const bool is_reduction_dim = src_dims[d] != dst_dims[d];
        if (is_reduction_dim) {
            reduce_dims[d] = src_dims[d];
            reduce_size *= reduce_dims[d];
        }
    }

    parallel_nd(idle_size, [= COMPAT_THIS_CAPTURE](dim_t l_offset) {
        dims_t idle_pos;
        utils::l_dims_by_l_offset(idle_pos, l_offset, dst_mdw.dims(), ndims);

        float acc_f32 = 0.f;
        if (types::is_integral_dt(acc_type)) {
            int acc = ker<int>(l_offset, alg, p, reduce_size, idle_pos,
                    reduce_dims, pd()->src_md(), src);
            acc_f32 = static_cast<float>(acc);
        } else {
            acc_f32 = ker<float>(l_offset, alg, p, reduce_size, idle_pos,
                    reduce_dims, pd()->src_md(), src);
        }

        finalize(acc_f32, alg, p, eps, reduce_size);

        const dim_t dst_off = dst_mdw.off_v(idle_pos);
        ref_post_ops_t::args_t args;
        args.dst_val = io::load_float_value(dst_mdw.data_type(), dst, dst_off);
        args.ctx = &ctx;
        args.l_offset = l_offset;
        args.dst_md = pd()->dst_md();
        ref_post_ops->execute(acc_f32, args);

        io::store_float_value(dst_mdw.data_type(), acc_f32, dst, dst_off);
    });

    return status::success;
}

} // namespace cpu
} // namespace impl
} // namespace dnnl