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
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
/*******************************************************************************
* Copyright 2018 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 COMMON_RNN_PD_HPP
#define COMMON_RNN_PD_HPP

#include "oneapi/dnnl/dnnl.h"

#include "c_types_map.hpp"
#include "primitive_desc.hpp"
#include "rnn.hpp"
#include "type_helpers.hpp"
#include "utils.hpp"

#define VDISPATCH_RNN(cond, msg, ...) \
    VCONDCHECK(primitive, create, dispatch, rnn, (cond), \
            status::unimplemented, "%s," msg, this->info(engine), \
            ##__VA_ARGS__)

#define VDISPATCH_RNN_SC(f, msg, ...) \
    VCHECK(primitive, create, dispatch, rnn, (f), "%s," msg, \
            this->info(engine), ##__VA_ARGS__)

namespace dnnl {
namespace impl {

struct rnn_fwd_pd_t;

// NOLINTBEGIN(google-default-arguments)
struct rnn_pd_t : public primitive_desc_t {
    static constexpr auto base_pkind = primitive_kind::rnn;

    const rnn_desc_t *desc() const { return &desc_; }
    const op_desc_t *op_desc() const override {
        return reinterpret_cast<const op_desc_t *>(this->desc());
    }

    status_t query(query_t what, int idx, void *result) const override {
        switch (what) {
            case query::prop_kind:
                *(prop_kind_t *)result = desc()->prop_kind;
                break;
            case query::cell_kind:
                *(alg_kind_t *)result = desc()->cell_kind;
                break;
            case query::activation_kind:
                *(alg_kind_t *)result = desc()->activation_kind;
                break;
            case query::direction:
                *(rnn_direction_t *)result = desc()->direction;
                break;
            case query::alpha_f32: *(float *)result = desc()->alpha; break;
            default: return primitive_desc_t::query(what, idx, result);
        }
        return status::success;
    }

    const memory_desc_t *src_md(
            int index = 0, bool user_input = false) const override {
        if (index == 0)
            return user_input ? &desc()->src_layer_desc : &src_layer_md_;
        if (index == 1 && with_src_iter())
            return user_input ? &desc()->src_iter_desc : &src_iter_md_;
        if (index == 2 && with_src_iter_c())
            return user_input ? &desc()->src_iter_c_desc : &src_iter_c_md_;
        return &glob_zero_md;
    }

    memory_desc_t &augru_attention_md() {
        if (with_augru_attention()) return weights_peephole_md_;
        return glob_zero_md;
    }

    const memory_desc_t &const_augru_attention_md() const {
        if (with_augru_attention()) return weights_peephole_md_;
        return glob_zero_md;
    }

    const memory_desc_t *weights_md(
            int index = 0, bool user_input = false) const override {
        if (index == 0)
            return user_input ? &desc()->weights_layer_desc
                              : &weights_layer_md_;
        if (index == 1)
            return user_input ? &desc()->weights_iter_desc : &weights_iter_md_;

        const int peephole_index = 2;
        if (is_lstm_peephole() && index == peephole_index)
            return user_input ? &desc()->weights_peephole_desc
                              : &weights_peephole_md_;

        const int projection_index = 2 + is_lstm_peephole();
        if (is_lstm_projection() && index == projection_index)
            return user_input ? &desc()->weights_projection_desc
                              : &weights_projection_md_;

        const int bias_index = 2 + is_lstm_peephole() + is_lstm_projection();
        if (with_bias() && index == bias_index)
            return user_input ? &desc()->bias_desc : &bias_md_;

        return &glob_zero_md;
    }
    const memory_desc_t *dst_md(
            int index = 0, bool user_input = false) const override {
        if (index == 0)
            return user_input ? &desc()->dst_layer_desc : &dst_layer_md_;
        if (index == 1 && with_dst_iter())
            return user_input ? &desc()->dst_iter_desc : &dst_iter_md_;
        if (index == 2 && with_dst_iter_c())
            return user_input ? &desc()->dst_iter_c_desc : &dst_iter_c_md_;
        return &glob_zero_md;
    }
    const memory_desc_t *workspace_md(int index = 0) const override {
        return (index == 0) ? &ws_md_ : &glob_zero_md;
    }

    /* common aux functions */

    bool is_training() const {
        return utils::one_of(desc_.prop_kind, prop_kind::forward_training,
                prop_kind::backward);
    }

    bool is_fwd() const {
        return utils::one_of(desc_.prop_kind, prop_kind::forward_training,
                prop_kind::forward_inference);
    }

    dim_t T() const { return desc_.src_layer_desc.dims[0]; }
    dim_t MB() const { return desc_.src_layer_desc.dims[1]; }

    dim_t L() const { return desc_.weights_layer_desc.dims[0]; }
    dim_t D() const { return desc_.weights_layer_desc.dims[1]; }

    dim_t SIC() const { return desc_.weights_iter_desc.dims[2]; }

    dim_t SLC() const { return desc_.weights_layer_desc.dims[2]; }
    dim_t G() const { return desc_.weights_layer_desc.dims[3]; }
    dim_t DHC() const { return desc_.weights_layer_desc.dims[4]; }

    // Returns the number of channels for the iter tensor.
    // Must be equal to the dst_iter.dims[3] if dst_iter is not zero.
    dim_t DIC() const {
        return is_lstm_projection() ? desc_.weights_projection_desc.dims[3]
                                    : DHC();
    }

    dim_t DLC() const { return desc_.dst_layer_desc.dims[2]; }

    bool with_bias() const {
        return !memory_desc_wrapper(desc_.bias_desc).is_zero();
    }

    bool with_augru_attention() const { return is_augru(); }

    bool with_src_iter() const {
        return !(memory_desc_wrapper(desc_.src_iter_desc).is_zero());
    }

    bool with_src_iter_c() const {
        return is_lstm()
                && !(memory_desc_wrapper(desc_.src_iter_desc).is_zero());
    }

    bool with_dst_iter() const {
        return !memory_desc_wrapper(desc_.dst_iter_desc).is_zero();
    }

    bool with_dst_iter_c() const {
        return is_lstm() && !memory_desc_wrapper(desc_.dst_iter_desc).is_zero();
    }

    dnnl::impl::alg_kind_t cell_kind() const { return desc_.cell_kind; }
    dnnl::impl::alg_kind_t activation_kind() const {
        return desc_.activation_kind;
    }

    bool is_lbr() const {
        return utils::one_of(cell_kind(), dnnl_lbr_gru, dnnl_lbr_augru);
    }

    bool is_augru() const {
        return utils::one_of(cell_kind(), dnnl_vanilla_augru, dnnl_lbr_augru);
    }

    bool is_lstm() const { return cell_kind() == dnnl_vanilla_lstm; }

    bool is_lstm_peephole() const {
        return is_lstm()
                && !memory_desc_wrapper(weights_peephole_md_).is_zero();
    }

    bool is_lstm_projection() const {
        return !memory_desc_wrapper(weights_projection_md_).is_zero();
    }

    bool diff_weights_overwrite() const {
        return desc_.flags & rnn_flags::diff_weights_overwrite;
    }

    dnnl_rnn_direction_t direction() const { return desc_.direction; }

protected:
    rnn_desc_t desc_;
    const rnn_fwd_pd_t *hint_fwd_pd_;

    memory_desc_t src_layer_md_;
    memory_desc_t src_iter_md_;
    memory_desc_t src_iter_c_md_;
    memory_desc_t weights_layer_md_;
    memory_desc_t weights_iter_md_;
    memory_desc_t weights_peephole_md_;
    memory_desc_t weights_projection_md_;
    memory_desc_t bias_md_;
    memory_desc_t dst_layer_md_;
    memory_desc_t dst_iter_md_;
    memory_desc_t dst_iter_c_md_;

    memory_desc_t ws_md_;

    rnn_pd_t(const op_desc_t *adesc, const primitive_attr_t *attr,
            const rnn_fwd_pd_t *hint_fwd_pd)
        : primitive_desc_t(attr, base_pkind)
        , desc_(*op_desc_t::to_desc<rnn_desc_t>(adesc))
        , hint_fwd_pd_(hint_fwd_pd)
        , src_layer_md_(desc_.src_layer_desc)
        , src_iter_md_(desc_.src_iter_desc)
        , src_iter_c_md_(desc_.src_iter_c_desc)
        , weights_layer_md_(desc_.weights_layer_desc)
        , weights_iter_md_(desc_.weights_iter_desc)
        , weights_peephole_md_(desc_.weights_peephole_desc)
        , weights_projection_md_(desc_.weights_projection_desc)
        , bias_md_(desc_.bias_desc)
        , dst_layer_md_(desc_.dst_layer_desc)
        , dst_iter_md_(desc_.dst_iter_desc)
        , dst_iter_c_md_(desc_.dst_iter_c_desc) {}
};
// NOLINTEND(google-default-arguments)

// NOLINTBEGIN(google-default-arguments)
struct rnn_fwd_pd_t : public rnn_pd_t {
    using base_class = rnn_fwd_pd_t;
    using hint_class = rnn_fwd_pd_t;

    arg_usage_t arg_usage(int arg) const override {
        if (arg == DNNL_ARG_SRC_LAYER) return arg_usage_t::input;

        if (arg == DNNL_ARG_AUGRU_ATTENTION)
            return with_augru_attention() ? arg_usage_t::input
                                          : arg_usage_t::unused;

        if (arg == DNNL_ARG_SRC_ITER)
            return with_src_iter() ? arg_usage_t::input : arg_usage_t::unused;

        if (arg == DNNL_ARG_SRC_ITER_C)
            return with_src_iter_c() ? arg_usage_t::input : arg_usage_t::unused;

        if (utils::one_of(arg, DNNL_ARG_WEIGHTS_LAYER, DNNL_ARG_WEIGHTS_ITER))
            return arg_usage_t::input;

        if (arg == DNNL_ARG_WEIGHTS_PEEPHOLE)
            return is_lstm_peephole() ? arg_usage_t::input
                                      : arg_usage_t::unused;

        if (arg == DNNL_ARG_WEIGHTS_PROJECTION)
            return is_lstm_projection() ? arg_usage_t::input
                                        : arg_usage_t::unused;

        if (arg == DNNL_ARG_BIAS)
            return with_bias() ? arg_usage_t::input : arg_usage_t::unused;

        if (arg == DNNL_ARG_DST_LAYER) return arg_usage_t::output;

        if (arg == DNNL_ARG_DST_ITER)
            return with_dst_iter() ? arg_usage_t::output : arg_usage_t::unused;

        if (arg == DNNL_ARG_DST_ITER_C)
            return with_dst_iter_c() ? arg_usage_t::output
                                     : arg_usage_t::unused;

        if (arg == DNNL_ARG_WORKSPACE)
            return is_training() ? arg_usage_t::output : arg_usage_t::unused;

        return primitive_desc_t::arg_usage(arg);
    }

    const memory_desc_t *arg_md(
            int arg, bool user_input = false) const override {
        switch (arg) {
            case DNNL_ARG_SRC_LAYER: return src_md(0);
            case DNNL_ARG_AUGRU_ATTENTION: return &const_augru_attention_md();
            case DNNL_ARG_SRC_ITER: return src_md(1);
            case DNNL_ARG_SRC_ITER_C: return src_md(2);
            case DNNL_ARG_WEIGHTS_LAYER: return weights_md(0);
            case DNNL_ARG_WEIGHTS_ITER: return weights_md(1);
            case DNNL_ARG_WEIGHTS_PEEPHOLE:
                return is_lstm_peephole() ? weights_md(2) : &glob_zero_md;
            case DNNL_ARG_WEIGHTS_PROJECTION:
                return is_lstm_projection() ? weights_md(2 + is_lstm_peephole())
                                            : &glob_zero_md;
            case DNNL_ARG_BIAS:
                return weights_md(
                        2 + is_lstm_peephole() + is_lstm_projection());
            case DNNL_ARG_DST_LAYER: return dst_md(0);
            case DNNL_ARG_DST_ITER: return dst_md(1);
            case DNNL_ARG_DST_ITER_C: return dst_md(2);
            default: return rnn_pd_t::arg_md(arg);
        }
    }

    int n_inputs() const override {
        return 3 + is_lstm_peephole() + is_lstm_projection() + with_bias()
                + with_src_iter() + with_src_iter_c() + is_augru();
    }
    int n_outputs() const override {
        return 1 + with_dst_iter() + with_dst_iter_c() + is_training();
    }

protected:
    rnn_fwd_pd_t(const op_desc_t *adesc, const primitive_attr_t *attr,
            const rnn_fwd_pd_t *hint_fwd_pd)
        : rnn_pd_t(adesc, attr, hint_fwd_pd) {}
};
// NOLINTEND(google-default-arguments)

// NOLINTBEGIN(google-default-arguments)
struct rnn_bwd_pd_t : public rnn_pd_t {
    using base_class = rnn_bwd_pd_t;
    using hint_class = rnn_fwd_pd_t;

    arg_usage_t arg_usage(int arg) const override {
        if (utils::one_of(arg, DNNL_ARG_SRC_LAYER, DNNL_ARG_DST_LAYER,
                    DNNL_ARG_DIFF_DST_LAYER, DNNL_ARG_WEIGHTS_LAYER,
                    DNNL_ARG_WEIGHTS_ITER))
            return arg_usage_t::input;

        if (utils::one_of(arg, DNNL_ARG_DIFF_SRC_LAYER,
                    DNNL_ARG_DIFF_WEIGHTS_LAYER, DNNL_ARG_DIFF_WEIGHTS_ITER))
            return arg_usage_t::output;

        if (arg == DNNL_ARG_AUGRU_ATTENTION)
            return with_augru_attention() ? arg_usage_t::input
                                          : arg_usage_t::unused;
        if (arg == DNNL_ARG_DIFF_AUGRU_ATTENTION)
            return with_augru_attention() ? arg_usage_t::output
                                          : arg_usage_t::unused;

        if (arg == DNNL_ARG_WEIGHTS_PEEPHOLE)
            return is_lstm_peephole() ? arg_usage_t::input
                                      : arg_usage_t::unused;
        if (arg == DNNL_ARG_DIFF_WEIGHTS_PEEPHOLE)
            return is_lstm_peephole() ? arg_usage_t::output
                                      : arg_usage_t::unused;

        if (arg == DNNL_ARG_WEIGHTS_PROJECTION)
            return is_lstm_projection() ? arg_usage_t::input
                                        : arg_usage_t::unused;
        if (arg == DNNL_ARG_DIFF_WEIGHTS_PROJECTION)
            return is_lstm_projection() ? arg_usage_t::output
                                        : arg_usage_t::unused;

        if (arg == DNNL_ARG_BIAS)
            return with_bias() ? arg_usage_t::input : arg_usage_t::unused;
        if (arg == DNNL_ARG_DIFF_BIAS)
            return with_bias() ? arg_usage_t::output : arg_usage_t::unused;

        if (arg == DNNL_ARG_SRC_ITER)
            return with_src_iter() ? arg_usage_t::input : arg_usage_t::unused;
        if (arg == DNNL_ARG_DIFF_SRC_ITER)
            return with_src_iter() ? arg_usage_t::output : arg_usage_t::unused;

        if (arg == DNNL_ARG_SRC_ITER_C)
            return with_src_iter_c() ? arg_usage_t::input : arg_usage_t::unused;
        if (arg == DNNL_ARG_DIFF_SRC_ITER_C)
            return with_src_iter_c() ? arg_usage_t::output
                                     : arg_usage_t::unused;

        if (arg == DNNL_ARG_DST_ITER)
            return with_dst_iter() ? arg_usage_t::input : arg_usage_t::unused;
        if (arg == DNNL_ARG_DIFF_DST_ITER)
            return with_dst_iter() ? arg_usage_t::input : arg_usage_t::unused;

        if (arg == DNNL_ARG_DST_ITER_C)
            return with_dst_iter_c() ? arg_usage_t::input : arg_usage_t::unused;
        if (arg == DNNL_ARG_DIFF_DST_ITER_C)
            return with_dst_iter_c() ? arg_usage_t::input : arg_usage_t::unused;

        if (arg == DNNL_ARG_WORKSPACE) return arg_usage_t::input;

        return primitive_desc_t::arg_usage(arg);
    }

    const memory_desc_t *arg_md(
            int arg, bool user_input = false) const override {
        switch (arg) {
            case DNNL_ARG_SRC_LAYER: return src_md(0);
            case DNNL_ARG_AUGRU_ATTENTION: return &const_augru_attention_md();
            case DNNL_ARG_SRC_ITER: return src_md(1);
            case DNNL_ARG_SRC_ITER_C: return src_md(2);
            case DNNL_ARG_DIFF_SRC_LAYER: return diff_src_md(0);
            case DNNL_ARG_DIFF_AUGRU_ATTENTION:
                return &const_diff_augru_attention_md();
            case DNNL_ARG_DIFF_SRC_ITER: return diff_src_md(1);
            case DNNL_ARG_DIFF_SRC_ITER_C: return diff_src_md(2);
            case DNNL_ARG_WEIGHTS_LAYER: return weights_md(0);
            case DNNL_ARG_WEIGHTS_ITER: return weights_md(1);
            case DNNL_ARG_WEIGHTS_PEEPHOLE:
                return is_lstm_peephole() ? weights_md(2) : &glob_zero_md;
            case DNNL_ARG_WEIGHTS_PROJECTION:
                return is_lstm_projection() ? weights_md(2 + is_lstm_peephole())
                                            : &glob_zero_md;
            case DNNL_ARG_BIAS:
                return weights_md(
                        2 + is_lstm_peephole() + is_lstm_projection());
            case DNNL_ARG_DIFF_WEIGHTS_LAYER: return diff_weights_md(0);
            case DNNL_ARG_DIFF_WEIGHTS_ITER: return diff_weights_md(1);
            case DNNL_ARG_DIFF_WEIGHTS_PEEPHOLE:
                return is_lstm_peephole() ? diff_weights_md(2) : &glob_zero_md;
            case DNNL_ARG_DIFF_WEIGHTS_PROJECTION:
                return is_lstm_projection()
                        ? diff_weights_md(2 + is_lstm_peephole())
                        : &glob_zero_md;
            case DNNL_ARG_DIFF_BIAS:
                return diff_weights_md(
                        2 + is_lstm_peephole() + is_lstm_projection());
            case DNNL_ARG_DST_LAYER: return dst_md(0);
            case DNNL_ARG_DST_ITER: return dst_md(1);
            case DNNL_ARG_DST_ITER_C: return dst_md(2);
            case DNNL_ARG_DIFF_DST_LAYER: return diff_dst_md(0);
            case DNNL_ARG_DIFF_DST_ITER: return diff_dst_md(1);
            case DNNL_ARG_DIFF_DST_ITER_C: return diff_dst_md(2);
            default: return rnn_pd_t::arg_md(arg);
        }
    }

    const memory_desc_t *diff_src_md(
            int index = 0, bool user_input = false) const override {
        if (index == 0)
            return user_input ? &desc()->diff_src_layer_desc
                              : &diff_src_layer_md_;
        if (index == 1 && with_src_iter())
            return user_input ? &desc()->diff_src_iter_desc
                              : &diff_src_iter_md_;
        if (index == 2 && with_src_iter_c())
            return user_input ? &desc()->diff_src_iter_c_desc
                              : &diff_src_iter_c_md_;
        return &glob_zero_md;
    }
    memory_desc_t &diff_augru_attention_md() {
        if (with_augru_attention()) return diff_weights_peephole_md_;
        return glob_zero_md;
    }
    const memory_desc_t &const_diff_augru_attention_md() const {
        if (with_augru_attention()) return diff_weights_peephole_md_;
        return glob_zero_md;
    }
    const memory_desc_t *diff_weights_md(
            int index = 0, bool user_input = false) const override {
        if (index == 0)
            return user_input ? &desc()->diff_weights_layer_desc
                              : &diff_weights_layer_md_;
        if (index == 1)
            return user_input ? &desc()->diff_weights_iter_desc
                              : &diff_weights_iter_md_;

        const int peephole_index = 2;
        if (is_lstm_peephole() && index == peephole_index)
            return user_input ? &desc()->diff_weights_peephole_desc
                              : &diff_weights_peephole_md_;

        const int projection_index = 2 + is_lstm_peephole();
        if (is_lstm_projection() && index == projection_index)
            return user_input ? &desc()->diff_weights_projection_desc
                              : &diff_weights_projection_md_;

        const int bias_index = 2 + is_lstm_peephole() + is_lstm_projection();
        if (with_bias() && index == bias_index)
            return user_input ? &desc()->diff_bias_desc : &diff_bias_md_;

        return &glob_zero_md;
    }
    const memory_desc_t *diff_dst_md(
            int index = 0, bool user_input = false) const override {
        if (index == 0)
            return user_input ? &desc()->diff_dst_layer_desc
                              : &diff_dst_layer_md_;
        if (index == 1 && with_dst_iter())
            return user_input ? &desc()->diff_dst_iter_desc
                              : &diff_dst_iter_md_;
        if (index == 2 && with_dst_iter_c())
            return user_input ? &desc()->diff_dst_iter_c_desc
                              : &diff_dst_iter_c_md_;
        return &glob_zero_md;
    }

    int n_inputs() const override {
        return 6 + with_src_iter() + with_src_iter_c()
                + 2 * (with_dst_iter() + with_dst_iter_c()) + is_lstm_peephole()
                + is_lstm_projection() + with_bias() + is_augru();
    }
    int n_outputs() const override {
        return 3 + with_src_iter() + with_src_iter_c() + is_lstm_peephole()
                + is_lstm_projection() + with_bias() + is_augru();
    }

protected:
    memory_desc_t diff_src_layer_md_;
    memory_desc_t diff_src_iter_md_;
    memory_desc_t diff_src_iter_c_md_;
    memory_desc_t diff_weights_layer_md_;
    memory_desc_t diff_weights_iter_md_;
    memory_desc_t diff_weights_peephole_md_;
    memory_desc_t diff_weights_projection_md_;
    memory_desc_t diff_bias_md_;
    memory_desc_t diff_dst_layer_md_;
    memory_desc_t diff_dst_iter_md_;
    memory_desc_t diff_dst_iter_c_md_;

    rnn_bwd_pd_t(const op_desc_t *adesc, const primitive_attr_t *attr,
            const rnn_fwd_pd_t *hint_fwd_pd)
        : rnn_pd_t(adesc, attr, hint_fwd_pd)
        , diff_src_layer_md_(desc_.diff_src_layer_desc)
        , diff_src_iter_md_(desc_.diff_src_iter_desc)
        , diff_src_iter_c_md_(desc_.diff_src_iter_c_desc)
        , diff_weights_layer_md_(desc_.diff_weights_layer_desc)
        , diff_weights_iter_md_(desc_.diff_weights_iter_desc)
        , diff_weights_peephole_md_(desc_.diff_weights_peephole_desc)
        , diff_weights_projection_md_(desc_.diff_weights_projection_desc)
        , diff_bias_md_(desc_.diff_bias_desc)
        , diff_dst_layer_md_(desc_.diff_dst_layer_desc)
        , diff_dst_iter_md_(desc_.diff_dst_iter_desc)
        , diff_dst_iter_c_md_(desc_.diff_dst_iter_c_desc) {}
};
// NOLINTEND(google-default-arguments)

} // namespace impl
} // namespace dnnl

#endif