distributed_py.cc 56.1 KB
Newer Older
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24
/* Copyright (c) 2022 PaddlePaddle Authors. All Rights Reserved.

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 <fcntl.h>
#ifdef _POSIX_C_SOURCE
#undef _POSIX_C_SOURCE
#endif

#ifdef _XOPEN_SOURCE
#undef _XOPEN_SOURCE
#endif

#include "paddle/fluid/distributed/collective/ProcessGroup.h"
25
#include "paddle/fluid/distributed/collective/ProcessGroupStream.h"
26
#include "paddle/fluid/distributed/collective/Types.h"
27
#include "paddle/fluid/distributed/collective/reducer.h"
28 29 30 31 32
#include "paddle/fluid/framework/lod_tensor.h"
#include "paddle/fluid/framework/tensor.h"
#include "paddle/fluid/imperative/layer.h"
#include "paddle/fluid/pybind/distributed_py.h"
#include "paddle/fluid/pybind/eager_utils.h"
33
#include "paddle/fluid/pybind/process_group_utils.h"
34 35
#include "paddle/phi/api/all.h"

36
#if defined(PADDLE_WITH_NCCL) || defined(PADDLE_WITH_RCCL)
37 38 39
#include "paddle/fluid/distributed/collective/ProcessGroupNCCL.h"
#endif

W
wuhuachaocoding 已提交
40 41 42 43
#if defined(PADDLE_WITH_MPI)
#include "paddle/fluid/distributed/collective/ProcessGroupMPI.h"
#endif

44 45 46 47
#if defined(PADDLE_WITH_CUSTOM_DEVICE)
#include "paddle/fluid/distributed/collective/ProcessGroupCustom.h"
#endif

48 49 50 51 52
#if defined(PADDLE_WITH_GLOO)
#include "paddle/fluid/distributed/collective/ProcessGroupGloo.h"
#include "paddle/fluid/distributed/store/tcp_store.h"
#endif

J
james 已提交
53 54 55 56
#if defined(PADDLE_WITH_XPU_BKCL)
#include "paddle/fluid/distributed/collective/ProcessGroupBKCL.h"
#endif

57 58
#include "paddle/phi/kernels/sync_batch_norm_kernel.h"

59 60 61 62 63 64 65
namespace py = pybind11;

namespace paddle {
namespace pybind {

using Tensor = paddle::experimental::Tensor;

66 67 68 69 70
std::shared_ptr<distributed::EagerReducer> CreateEagerReducer(
    py::handle py_tensors,
    const std::vector<std::vector<size_t>> &group_indices,
    const std::vector<bool> &is_sparse_gradient,
    std::shared_ptr<distributed::ProcessGroup> process_group,
71 72
    const std::vector<size_t> &group_size_limits,
    bool find_unused_parameters) {
73
  auto params = CastPyArg2VectorOfTensor(py_tensors.ptr(), 0);
74 75 76 77 78 79
  return std::make_shared<distributed::EagerReducer>(params,
                                                     group_indices,
                                                     is_sparse_gradient,
                                                     process_group,
                                                     group_size_limits,
                                                     find_unused_parameters);
80 81
}

82 83 84 85 86 87 88 89
#if defined(PADDLE_WITH_GLOO)
using ProcessGroupGloo = paddle::distributed::ProcessGroupGloo;
using GlooStore = paddle::distributed::ProcessGroupGloo::GlooStore;
using GlooOptions = paddle::distributed::ProcessGroupGloo::GlooOptions;
#endif

static std::string GLOO_SOCKET_IFNAME_ENV = "GLOO_SOCKET_IFNAME";  // NOLINT

90 91 92
static UNUSED void *use_ccl_comm_func =
    phi::detail::GetCCLComm(phi::CPUPlace());

93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110
void BindDistributed(py::module *m) {
  py::enum_<distributed::ReduceOp>(*m, "ReduceOp")
      .value("SUM", distributed::ReduceOp::SUM)
      .value("AVG", distributed::ReduceOp::AVG)
      .value("MAX", distributed::ReduceOp::MAX)
      .value("MIN", distributed::ReduceOp::MIN)
      .value("PRODUCT", distributed::ReduceOp::PRODUCT);

  py::class_<distributed::AllreduceOptions>(*m, "AllreduceOptions")
      .def(py::init<>())
      .def_readwrite("reduce_op", &distributed::AllreduceOptions::reduce_op);

  py::class_<distributed::BroadcastOptions>(*m, "BroadcastOptions")
      .def(py::init<>())
      .def_readwrite("source_rank", &distributed::BroadcastOptions::source_rank)
      .def_readwrite("source_root",
                     &distributed::BroadcastOptions::source_root);

B
Baibaifan 已提交
111 112
  py::class_<distributed::BarrierOptions>(*m, "BarrierOptions")
      .def(py::init<>())
113
      .def_readwrite("device_id", &distributed::BarrierOptions::device_id);
B
Baibaifan 已提交
114

115 116 117 118 119
  py::class_<distributed::ReduceOptions>(*m, "ReduceOptions")
      .def(py::init<>())
      .def_readwrite("reduce_op", &distributed::ReduceOptions::reduce_op)
      .def_readwrite("source_root", &distributed::ReduceOptions::root_rank);

120 121 122 123 124 125
  auto ProcessGroup =
      py::class_<distributed::ProcessGroup,
                 std::shared_ptr<distributed::ProcessGroup>>(*m, "ProcessGroup")
          .def("rank", &distributed::ProcessGroup::GetRank)
          .def("size", &distributed::ProcessGroup::GetSize)
          .def("name", &distributed::ProcessGroup::GetBackendName)
126
          .def(
L
LiYuRio 已提交
127
              "all_reduce",
128 129 130 131 132
              [](distributed::ProcessGroup &self,
                 py::handle py_tensor,
                 distributed::ReduceOp op,
                 bool sync_op) {
                auto tensor = CastPyArg2Tensor(py_tensor.ptr(), 0);
133
                auto p_dense =
134
                    std::dynamic_pointer_cast<phi::DenseTensor>(tensor.impl());
135 136 137 138
                auto *out_dense = p_dense.get();
                auto in_dense = *p_dense;
                distributed::AllreduceOptions opts{op};
                return self.AllReduce(out_dense, in_dense, opts, sync_op);
139 140 141 142 143 144
              },
              py::arg("tensor"),
              py::arg("op"),
              py::arg("sync_op"),
              py::call_guard<py::gil_scoped_release>())

145 146 147 148 149 150 151
          .def(
              "broadcast",
              [](distributed::ProcessGroup &self,
                 py::handle py_tensor,
                 int src,
                 bool sync_op) {
                auto tensor = CastPyArg2Tensor(py_tensor.ptr(), 0);
152
                auto p_dense =
153
                    std::dynamic_pointer_cast<phi::DenseTensor>(tensor.impl());
154 155 156 157
                auto *out_dense = p_dense.get();
                auto in_dense = *p_dense;
                distributed::BroadcastOptions opts{src};
                return self.Broadcast(out_dense, in_dense, opts, sync_op);
158 159 160 161 162 163
              },
              py::arg("tensor"),
              py::arg("src"),
              py::arg("sync_op"),
              py::call_guard<py::gil_scoped_release>())

164 165 166 167 168 169 170
          .def(
              "send",
              [](distributed::ProcessGroup &self,
                 py::handle py_tensor,
                 int dst,
                 bool sync_op) {
                auto tensor = CastPyArg2Tensor(py_tensor.ptr(), 0);
171
                auto p_dense =
172
                    std::dynamic_pointer_cast<phi::DenseTensor>(tensor.impl());
173
                auto *out_dense = p_dense.get();
174 175 176
                // numel == -1 indicates sending the whole tensor
                return self.Send(
                    out_dense, dst, /*offset*/ 0, /*numel*/ -1, sync_op);
177 178 179 180 181 182 183 184 185 186 187 188 189 190 191
              },
              py::arg("tensor"),
              py::arg("dst"),
              py::arg("sync_op"),
              py::call_guard<py::gil_scoped_release>())

          .def(
              "send_partial",
              [](distributed::ProcessGroup &self,
                 py::handle py_tensor,
                 int dst_rank,
                 int nranks,
                 int rank_id,
                 bool sync_op) {
                auto tensor = CastPyArg2Tensor(py_tensor.ptr(), 0);
192
                auto p_dense =
193
                    std::dynamic_pointer_cast<phi::DenseTensor>(tensor.impl());
194 195
                auto *out_dense = p_dense.get();

196
                int64_t numel = p_dense->numel();
197 198
                int64_t send_numel = numel / nranks;
                int64_t offset = send_numel * rank_id;
199 200

                return self.Send(
201
                    out_dense, dst_rank, offset, send_numel, sync_op);
202 203 204 205 206
              },
              py::arg("tensor"),
              py::arg("dst"),
              py::arg("num"),
              py::arg("id"),
207
              py::arg("sync_op") = true,
208 209 210 211 212 213 214 215 216
              py::call_guard<py::gil_scoped_release>())

          .def(
              "recv",
              [](distributed::ProcessGroup &self,
                 py::handle py_tensor,
                 int src,
                 bool sync_op) {
                auto tensor = CastPyArg2Tensor(py_tensor.ptr(), 0);
217
                auto p_dense =
218
                    std::dynamic_pointer_cast<phi::DenseTensor>(tensor.impl());
219
                auto *in_dense = p_dense.get();
220 221 222
                // numel == -1 indicates receiving the whole tensor
                return self.Recv(
                    in_dense, src, /*offset*/ 0, /*numel*/ -1, sync_op);
223 224 225 226 227 228 229 230 231 232 233 234 235 236 237
              },
              py::arg("tensor"),
              py::arg("src"),
              py::arg("sync_op"),
              py::call_guard<py::gil_scoped_release>())

          .def(
              "recv_partial",
              [](distributed::ProcessGroup &self,
                 py::handle py_tensor,
                 int src_rank,
                 int nranks,
                 int rank_id,
                 bool sync_op) {
                auto tensor = CastPyArg2Tensor(py_tensor.ptr(), 0);
238
                auto p_dense =
239
                    std::dynamic_pointer_cast<phi::DenseTensor>(tensor.impl());
240 241
                auto *out_dense = p_dense.get();

242
                int64_t numel = p_dense->numel();
243 244
                int64_t recv_numel = numel / nranks;
                int64_t offset = recv_numel * rank_id;
245 246

                return self.Recv(
247
                    out_dense, src_rank, offset, recv_numel, sync_op);
248 249 250 251 252
              },
              py::arg("tensor"),
              py::arg("src"),
              py::arg("num"),
              py::arg("id"),
253
              py::arg("sync_op") = true,
254 255
              py::call_guard<py::gil_scoped_release>())

256 257
          .def(
              "all_gather",
258 259
              [](distributed::ProcessGroup &self,
                 py::handle py_out_tensor_list,
260
                 py::handle py_in_tensor,
261 262 263 264
                 bool sync_op) {
                auto out_tensor_list =
                    CastPyArg2VectorOfTensor(py_out_tensor_list.ptr(), 0);
                Tensor concat_out_tensor = paddle::concat(out_tensor_list, 0);
265
                auto p_out_tensor = std::dynamic_pointer_cast<phi::DenseTensor>(
266
                    concat_out_tensor.impl());
267 268 269 270 271 272
                auto *out_dense = p_out_tensor.get();

                auto in_tensor = CastPyArg2Tensor(py_in_tensor.ptr(), 0);
                auto p_in_tensor = std::dynamic_pointer_cast<phi::DenseTensor>(
                    in_tensor.impl());
                auto in_dense = *p_in_tensor;
273

274
                const auto &dev_ctx = self.GetDeviceContext(in_tensor.place());
275 276 277 278 279
                auto task = self.AllGather(out_dense,
                                           in_dense,
                                           /*offset*/ 0,
                                           /*numel*/ -1,
                                           sync_op);
280
                distributed::SplitTensor(dev_ctx, *out_dense, &out_tensor_list);
281
                task->UpdateWaitChain(dev_ctx);
282 283 284
                return task;
              },
              py::arg("out"),
285
              py::arg("in"),
286 287 288 289
              py::arg("sync_op"),
              py::call_guard<py::gil_scoped_release>())

          .def(
L
LiYuRio 已提交
290
              "all_gather_into_tensor",
291 292
              [](distributed::ProcessGroup &self,
                 py::handle py_out_tensor,
293
                 py::handle py_in_tensor,
294 295
                 bool sync_op) {
                auto out_tensor = CastPyArg2Tensor(py_out_tensor.ptr(), 0);
296
                auto p_out_tensor = std::dynamic_pointer_cast<phi::DenseTensor>(
297
                    out_tensor.impl());
298 299 300 301 302 303
                auto *out_dense = p_out_tensor.get();

                auto in_tensor = CastPyArg2Tensor(py_in_tensor.ptr(), 0);
                auto p_in_tensor = std::dynamic_pointer_cast<phi::DenseTensor>(
                    in_tensor.impl());
                auto in_dense = *p_in_tensor;
304

305 306 307 308 309
                return self.AllGather(out_dense,
                                      in_dense,
                                      /*offset*/ 0,
                                      /*numel*/ -1,
                                      sync_op);
310 311
              },
              py::arg("out"),
312
              py::arg("in"),
313 314 315
              py::arg("sync_op"),
              py::call_guard<py::gil_scoped_release>())

316
          .def(
L
LiYuRio 已提交
317
              "all_to_all",
318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336
              [](distributed::ProcessGroup &self,
                 py::handle py_in_tensor_list,
                 py::handle py_out_tensor_list,
                 bool sync_op) {
                auto in_tensor_list =
                    CastPyArg2VectorOfTensor(py_in_tensor_list.ptr(), 0);
                Tensor concat_in_tensor = paddle::concat(in_tensor_list, 0);
                auto in_dense = std::dynamic_pointer_cast<phi::DenseTensor>(
                    concat_in_tensor.impl());
                std::vector<phi::DenseTensor> in_wrapper = {*in_dense};

                auto out_tensor_list =
                    CastPyArg2VectorOfTensor(py_out_tensor_list.ptr(), 0);
                Tensor concat_out_tensor = paddle::concat(out_tensor_list, 0);
                auto out_dense = std::dynamic_pointer_cast<phi::DenseTensor>(
                    concat_out_tensor.impl());
                std::vector<phi::DenseTensor> out_wrapper = {*out_dense};

                // in_tensor_list should not be empty
337
                const auto &dev_ctx =
338 339 340
                    self.GetDeviceContext(in_tensor_list.back().place());
                auto task = self.AllToAll(in_wrapper, out_wrapper, sync_op);
                distributed::SplitTensor(dev_ctx, *out_dense, &out_tensor_list);
341
                task->UpdateWaitChain(dev_ctx);
342 343 344 345 346 347 348 349
                return task;
              },
              py::arg("in"),
              py::arg("out"),
              py::arg("sync_op"),
              py::call_guard<py::gil_scoped_release>())

          .def(
L
LiYuRio 已提交
350
              "all_to_all_tensor",
351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371
              [](distributed::ProcessGroup &self,
                 py::handle py_in_tensor,
                 py::handle py_out_tensor,
                 bool sync_op) {
                auto in_tensor = CastPyArg2Tensor(py_in_tensor.ptr(), 0);
                auto in_dense = std::dynamic_pointer_cast<phi::DenseTensor>(
                    in_tensor.impl());
                std::vector<phi::DenseTensor> in_wrapper = {*in_dense};

                auto out_tensor = CastPyArg2Tensor(py_out_tensor.ptr(), 0);
                auto out_dense = std::dynamic_pointer_cast<phi::DenseTensor>(
                    out_tensor.impl());
                std::vector<phi::DenseTensor> out_wrapper = {*out_dense};

                return self.AllToAll(in_wrapper, out_wrapper, sync_op);
              },
              py::arg("in"),
              py::arg("out"),
              py::arg("sync_op"),
              py::call_guard<py::gil_scoped_release>())

372
          .def(
L
LiYuRio 已提交
373
              "all_to_all_single",
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
              [](distributed::ProcessGroup &self,
                 py::handle py_in_tensor,
                 py::handle py_out_tensor,
                 std::vector<int64_t> &in_sizes,
                 std::vector<int64_t> &out_sizes,
                 bool sync_op) {
                auto in_tensor = CastPyArg2Tensor(py_in_tensor.ptr(), 0);
                auto in_dense = std::dynamic_pointer_cast<phi::DenseTensor>(
                    in_tensor.impl());
                std::vector<phi::DenseTensor> in_wrapper = {*in_dense};

                auto out_tensor = CastPyArg2Tensor(py_out_tensor.ptr(), 0);
                auto out_dense = std::dynamic_pointer_cast<phi::DenseTensor>(
                    out_tensor.impl());
                std::vector<phi::DenseTensor> out_wrapper = {*out_dense};

                return self.AllToAllSingle(
                    in_wrapper, out_wrapper, in_sizes, out_sizes, sync_op);
              },
              py::arg("in"),
              py::arg("out"),
              py::arg("in_sizes"),
              py::arg("out_sizes"),
              py::arg("sync_op"),
              py::call_guard<py::gil_scoped_release>())

          .def(
              "reduce",
              [](distributed::ProcessGroup &self,
                 py::handle py_in_tensor,
                 int dst,
                 distributed::ReduceOp op,
                 bool sync_op) {
                auto in_tensor = CastPyArg2Tensor(py_in_tensor.ptr(), 0);
                distributed::ReduceOptions opts{op, dst};
                auto dense = std::dynamic_pointer_cast<phi::DenseTensor>(
                    in_tensor.impl());
                std::vector<phi::DenseTensor> tensors = {*dense};
                return self.Reduce(tensors, tensors, opts, sync_op);
              },
              py::arg("tensor"),
              py::arg("dst"),
              py::arg("op"),
              py::arg("sync_op"),
              py::call_guard<py::gil_scoped_release>())

          .def(
              "reduce_scatter",
              [](distributed::ProcessGroup &self,
                 py::handle py_in_tensor_list,
                 py::handle py_out_tensor,
                 distributed::ReduceOp op,
                 bool sync_op) {
                auto in_tensor_list =
                    CastPyArg2VectorOfTensor(py_in_tensor_list.ptr(), 0);
                Tensor concat_in_tensor = paddle::concat(in_tensor_list, 0);
                auto in_dense = std::dynamic_pointer_cast<phi::DenseTensor>(
                    concat_in_tensor.impl());
                std::vector<phi::DenseTensor> in_wrapper = {*in_dense};

                auto out_tensor = CastPyArg2Tensor(py_out_tensor.ptr(), 0);
                auto out_dense = std::dynamic_pointer_cast<phi::DenseTensor>(
                    out_tensor.impl());
                std::vector<phi::DenseTensor> out_wrapper = {*out_dense};

                distributed::ReduceScatterOptions opts{op};
                return self.ReduceScatter(
                    in_wrapper, out_wrapper, opts, sync_op);
              },
              py::arg("in"),
              py::arg("out"),
              py::arg("op"),
              py::arg("sync_op"),
              py::call_guard<py::gil_scoped_release>())

          .def(
              "reduce_scatter_tensor",
              [](distributed::ProcessGroup &self,
                 py::handle py_in_tensor,
                 py::handle py_out_tensor,
                 distributed::ReduceOp op,
                 bool sync_op) {
                auto in_tensor = CastPyArg2Tensor(py_in_tensor.ptr(), 0);
                auto in_dense = std::dynamic_pointer_cast<phi::DenseTensor>(
                    in_tensor.impl());
                std::vector<phi::DenseTensor> in_wrapper = {*in_dense};

                auto out_tensor = CastPyArg2Tensor(py_out_tensor.ptr(), 0);
                auto out_dense = std::dynamic_pointer_cast<phi::DenseTensor>(
                    out_tensor.impl());
                std::vector<phi::DenseTensor> out_wrapper = {*out_dense};

                distributed::ReduceScatterOptions opts{op};
                return self.ReduceScatter(
                    in_wrapper, out_wrapper, opts, sync_op);
              },
              py::arg("in"),
              py::arg("out"),
              py::arg("op"),
              py::arg("sync_op"),
              py::call_guard<py::gil_scoped_release>())

          .def(
              "scatter",
              [](distributed::ProcessGroup &self,
                 py::handle py_in_tensor_list,
                 py::handle py_out_tensor,
                 int src,
                 bool sync_op) {
                auto in_tensor_list =
                    CastPyArg2VectorOfTensor(py_in_tensor_list.ptr(), 0);
                Tensor concat_in_tensor = paddle::concat(in_tensor_list, 0);
                auto in_dense = std::dynamic_pointer_cast<phi::DenseTensor>(
                    concat_in_tensor.impl());
                std::vector<phi::DenseTensor> in_wrapper = {*in_dense};

                auto out_tensor = CastPyArg2Tensor(py_out_tensor.ptr(), 0);
                auto out_dense = std::dynamic_pointer_cast<phi::DenseTensor>(
                    out_tensor.impl());
                std::vector<phi::DenseTensor> out_wrapper = {*out_dense};

                distributed::ScatterOptions opts{src};
                return self.Scatter(in_wrapper, out_wrapper, opts, sync_op);
              },
              py::arg("in"),
              py::arg("out"),
              py::arg("src"),
              py::arg("sync_op"),
              py::call_guard<py::gil_scoped_release>())

          .def(
              "scatter_tensor",
              [](distributed::ProcessGroup &self,
                 py::handle py_in_tensor,
                 py::handle py_out_tensor,
                 int src,
                 bool sync_op) {
                auto in_tensor = CastPyArg2Tensor(py_in_tensor.ptr(), 0);
                auto in_dense = std::dynamic_pointer_cast<phi::DenseTensor>(
                    in_tensor.impl());
                std::vector<phi::DenseTensor> in_wrapper = {*in_dense};

                auto out_tensor = CastPyArg2Tensor(py_out_tensor.ptr(), 0);
                auto out_dense = std::dynamic_pointer_cast<phi::DenseTensor>(
                    out_tensor.impl());
                std::vector<phi::DenseTensor> out_wrapper = {*out_dense};

                distributed::ScatterOptions opts{src};
                return self.Scatter(in_wrapper, out_wrapper, opts, sync_op);
              },
              py::arg("in"),
              py::arg("out"),
              py::arg("src"),
              py::arg("sync_op"),
              py::call_guard<py::gil_scoped_release>())

L
LiYuRio 已提交
530 531
          .def(
              "barrier",
532
              [](distributed::ProcessGroup &self, int8_t device_id) {
L
LiYuRio 已提交
533
                distributed::BarrierOptions opts;
534
                opts.device_id = device_id;
L
LiYuRio 已提交
535 536
                return self.Barrier(opts);
              },
537
              py::arg("device_id") = -1,
L
LiYuRio 已提交
538 539 540 541 542 543 544 545 546 547 548 549 550 551 552 553 554 555 556 557 558 559 560 561 562 563 564 565 566 567 568 569 570 571 572 573 574 575 576 577 578 579 580 581 582 583 584 585 586 587 588 589 590 591 592 593 594 595 596 597 598 599 600 601 602 603 604 605 606 607 608 609 610 611 612 613 614 615 616 617 618 619 620 621 622 623 624 625 626 627
              py::call_guard<py::gil_scoped_release>())

          // TODO(liyurui): Interface below will be removed in the future.
          .def(
              "allreduce",
              [](distributed::ProcessGroup &self,
                 py::handle py_tensor,
                 distributed::ReduceOp op) {
                auto tensor = CastPyArg2Tensor(py_tensor.ptr(), 0);
                distributed::AllreduceOptions opts;
                opts.reduce_op = op;
                auto dense =
                    std::dynamic_pointer_cast<phi::DenseTensor>(tensor.impl());
                std::vector<phi::DenseTensor> tensors = {*dense};
                return self.AllReduce(tensors, tensors, opts);
              },
              py::arg("tensor"),
              py::arg("op") = distributed::ReduceOp::SUM,
              py::call_guard<py::gil_scoped_release>())

          .def(
              "broadcast",
              [](distributed::ProcessGroup &self,
                 py::handle py_tensor,
                 int source_rank) {
                auto tensor = CastPyArg2Tensor(py_tensor.ptr(), 0);
                distributed::BroadcastOptions opts;
                opts.source_rank = source_rank;
                auto dense =
                    std::dynamic_pointer_cast<phi::DenseTensor>(tensor.impl());
                std::vector<phi::DenseTensor> tensors = {*dense};
                return self.Broadcast(tensors, tensors, opts);
              },
              py::arg("tensor"),
              py::arg("source_rank"),
              py::call_guard<py::gil_scoped_release>())

          .def(
              "send",
              [](distributed::ProcessGroup &self,
                 py::handle py_tensor,
                 int dst) {
                auto tensor = CastPyArg2Tensor(py_tensor.ptr(), 0);
                auto dense =
                    std::dynamic_pointer_cast<phi::DenseTensor>(tensor.impl());
                std::vector<phi::DenseTensor> tensors = {*dense};
                return self.Send(tensors, dst);
              },
              py::arg("tensor"),
              py::arg("dst"),
              py::call_guard<py::gil_scoped_release>())

          .def(
              "recv",
              [](distributed::ProcessGroup &self,
                 py::handle py_tensor,
                 int src) {
                auto tensor = CastPyArg2Tensor(py_tensor.ptr(), 0);
                auto dense =
                    std::dynamic_pointer_cast<phi::DenseTensor>(tensor.impl());
                std::vector<phi::DenseTensor> tensors = {*dense};
                return self.Recv(tensors, src);
              },
              py::arg("tensor"),
              py::arg("src"),
              py::call_guard<py::gil_scoped_release>())

          .def(
              "all_gather",
              [](distributed::ProcessGroup &self,
                 py::handle py_in_tensor,
                 py::handle py_out_tensor) {
                auto in_tensor = CastPyArg2Tensor(py_in_tensor.ptr(), 0);
                auto out_tensor = CastPyArg2Tensor(py_out_tensor.ptr(), 0);
                auto in_dense = std::dynamic_pointer_cast<phi::DenseTensor>(
                    in_tensor.impl());
                auto out_dense = std::dynamic_pointer_cast<phi::DenseTensor>(
                    out_tensor.impl());
                std::vector<phi::DenseTensor> in_tensors = {*in_dense};
                std::vector<phi::DenseTensor> out_tensors = {*out_dense};
                return self.AllGather(in_tensors, out_tensors);
              },
              py::arg("in"),
              py::arg("out"),
              py::call_guard<py::gil_scoped_release>())

          .def(
              "all_gather_partial",
              [](distributed::ProcessGroup &self,
                 py::handle py_out_tensor,
628
                 py::handle py_in_tensor,
L
LiYuRio 已提交
629 630 631
                 int nranks,
                 int rank_id) {
                auto out_tensor = CastPyArg2Tensor(py_out_tensor.ptr(), 0);
632
                auto p_out_tensor = std::dynamic_pointer_cast<phi::DenseTensor>(
L
LiYuRio 已提交
633
                    out_tensor.impl());
634 635 636 637 638 639 640 641
                auto *out_dense = p_out_tensor.get();

                auto in_tensor = CastPyArg2Tensor(py_in_tensor.ptr(), 0);
                auto p_in_tensor = std::dynamic_pointer_cast<phi::DenseTensor>(
                    in_tensor.impl());
                auto in_dense = *p_in_tensor;

                int64_t numel = in_dense.numel();
L
LiYuRio 已提交
642 643
                int64_t send_numel = numel / nranks;
                int64_t offset = send_numel * rank_id;
644 645
                return self.AllGather(
                    out_dense, in_dense, offset, send_numel, /*sync_op*/ true);
L
LiYuRio 已提交
646 647
              },
              py::arg("out"),
648
              py::arg("in"),
L
LiYuRio 已提交
649 650 651 652 653 654 655 656 657 658 659 660 661 662 663 664 665 666 667 668 669 670 671 672 673 674 675 676 677 678 679 680 681 682 683 684 685 686 687 688 689 690 691 692 693 694 695 696 697 698 699 700 701 702 703 704 705 706 707 708 709 710 711 712 713 714 715 716 717 718 719 720 721 722 723 724 725 726 727 728 729 730 731 732 733 734 735 736
              py::arg("num"),
              py::arg("id"),
              py::call_guard<py::gil_scoped_release>())

          .def(
              "alltoall",
              [](distributed::ProcessGroup &self,
                 py::handle py_in_tensor,
                 py::handle py_out_tensor) {
                auto in_tensor = CastPyArg2Tensor(py_in_tensor.ptr(), 0);
                auto out_tensor = CastPyArg2Tensor(py_out_tensor.ptr(), 0);
                auto in_dense = std::dynamic_pointer_cast<phi::DenseTensor>(
                    in_tensor.impl());
                auto out_dense = std::dynamic_pointer_cast<phi::DenseTensor>(
                    out_tensor.impl());
                std::vector<phi::DenseTensor> in_tensors = {*in_dense};
                std::vector<phi::DenseTensor> out_tensors = {*out_dense};
                return self.AllToAll(in_tensors, out_tensors);
              },
              py::arg("in"),
              py::arg("out"),
              py::call_guard<py::gil_scoped_release>())

          .def(
              "alltoall_single",
              [](distributed::ProcessGroup &self,
                 py::handle py_in_tensor,
                 py::handle py_out_tensor,
                 std::vector<int64_t> in_sizes,
                 std::vector<int64_t> out_sizes) {
                auto in_tensor = CastPyArg2Tensor(py_in_tensor.ptr(), 0);
                auto out_tensor = CastPyArg2Tensor(py_out_tensor.ptr(), 0);
                auto in_dense = std::dynamic_pointer_cast<phi::DenseTensor>(
                    in_tensor.impl());
                auto out_dense = std::dynamic_pointer_cast<phi::DenseTensor>(
                    out_tensor.impl());
                std::vector<phi::DenseTensor> in_tensors = {*in_dense};
                std::vector<phi::DenseTensor> out_tensors = {*out_dense};
                return self.AllToAll_Single(
                    in_tensors, out_tensors, in_sizes, out_sizes);
              },
              py::arg("in"),
              py::arg("out"),
              py::arg("in_sizes"),
              py::arg("out_sizes"),
              py::call_guard<py::gil_scoped_release>())

          .def(
              "reduce",
              [](distributed::ProcessGroup &self,
                 py::handle py_in_tensor,
                 int dst,
                 distributed::ReduceOp op) {
                auto in_tensor = CastPyArg2Tensor(py_in_tensor.ptr(), 0);
                distributed::ReduceOptions opts;
                opts.reduce_op = op;
                opts.root_rank = dst;
                auto dense = std::dynamic_pointer_cast<phi::DenseTensor>(
                    in_tensor.impl());
                std::vector<phi::DenseTensor> tensors = {*dense};
                return self.Reduce(tensors, tensors, opts);
              },
              py::arg("tensor"),
              py::arg("dst"),
              py::arg("op") = distributed::ReduceOp::SUM,
              py::call_guard<py::gil_scoped_release>())

          .def(
              "scatter",
              [](distributed::ProcessGroup &self,
                 py::handle py_in_tensor,
                 py::handle py_out_tensor,
                 int src) {
                auto in_tensor = CastPyArg2Tensor(py_in_tensor.ptr(), 0);
                auto out_tensor = CastPyArg2Tensor(py_out_tensor.ptr(), 0);
                distributed::ScatterOptions opts;
                opts.root_rank = src;
                auto in_dense = std::dynamic_pointer_cast<phi::DenseTensor>(
                    in_tensor.impl());
                auto out_dense = std::dynamic_pointer_cast<phi::DenseTensor>(
                    out_tensor.impl());
                std::vector<phi::DenseTensor> in_tensors = {*in_dense};
                std::vector<phi::DenseTensor> out_tensors = {*out_dense};
                return self.Scatter(in_tensors, out_tensors, opts);
              },
              py::arg("in"),
              py::arg("out"),
              py::arg("src"),
737
              py::call_guard<py::gil_scoped_release>());
738

739 740 741 742
  auto ProcessGroupStream =
      py::class_<distributed::ProcessGroupStream,
                 std::shared_ptr<distributed::ProcessGroupStream>>(
          *m, "ProcessGroupStream", ProcessGroup)
743
          .def(
L
LiYuRio 已提交
744
              "all_gather_on_calc_stream",
745
              [](distributed::ProcessGroupStream &self,
746 747
                 py::handle py_out_tensor_list,
                 py::handle py_in_tensor) {
748 749 750
                auto out_tensor_list =
                    CastPyArg2VectorOfTensor(py_out_tensor_list.ptr(), 0);
                Tensor concat_out_tensor = paddle::concat(out_tensor_list, 0);
751
                auto p_out_tensor = std::dynamic_pointer_cast<phi::DenseTensor>(
752
                    concat_out_tensor.impl());
753 754 755 756 757 758
                auto *out_dense = p_out_tensor.get();

                auto in_tensor = CastPyArg2Tensor(py_in_tensor.ptr(), 0);
                auto p_in_tensor = std::dynamic_pointer_cast<phi::DenseTensor>(
                    in_tensor.impl());
                auto in_dense = *p_in_tensor;
759

760
                const auto &dev_ctx =
761
                    self.GetDeviceContext(in_tensor.place(), true);
762 763
                auto task = self.AllGather(out_dense,
                                           in_dense,
764 765
                                           /*offset*/ 0,
                                           /*numel*/ -1,
766 767 768 769 770 771
                                           /*sync_op*/ true,
                                           /*use_calc_stream*/ true);
                distributed::SplitTensor(dev_ctx, *out_dense, &out_tensor_list);
                return task;
              },
              py::arg("out"),
772
              py::arg("in"),
773 774 775
              py::call_guard<py::gil_scoped_release>())

          .def(
L
LiYuRio 已提交
776
              "all_gather_into_tensor_on_calc_stream",
777
              [](distributed::ProcessGroupStream &self,
778 779
                 py::handle py_out_tensor,
                 py::handle py_in_tensor) {
780
                auto out_tensor = CastPyArg2Tensor(py_out_tensor.ptr(), 0);
781
                auto p_out_tensor = std::dynamic_pointer_cast<phi::DenseTensor>(
782
                    out_tensor.impl());
783
                auto *out_dense = p_out_tensor.get();
784

785 786 787 788 789 790 791
                auto in_tensor = CastPyArg2Tensor(py_in_tensor.ptr(), 0);
                auto p_in_tensor = std::dynamic_pointer_cast<phi::DenseTensor>(
                    in_tensor.impl());
                auto in_dense = *p_in_tensor;

                return self.AllGather(out_dense,
                                      in_dense,
792 793
                                      /*offset*/ 0,
                                      /*numel*/ -1,
794 795 796 797
                                      /*sync_op*/ true,
                                      /*use_calc_stream*/ true);
              },
              py::arg("out"),
798
              py::arg("in"),
799 800
              py::call_guard<py::gil_scoped_release>())

801 802 803 804
          .def(
              "all_gather_partial_on_calc_stream",
              [](distributed::ProcessGroupStream &self,
                 py::handle py_out_tensor,
805
                 py::handle py_in_tensor,
806 807 808
                 int nranks,
                 int rank_id) {
                auto out_tensor = CastPyArg2Tensor(py_out_tensor.ptr(), 0);
809
                auto p_out_tensor = std::dynamic_pointer_cast<phi::DenseTensor>(
810
                    out_tensor.impl());
811 812 813 814 815 816 817 818
                auto *out_dense = p_out_tensor.get();

                auto in_tensor = CastPyArg2Tensor(py_in_tensor.ptr(), 0);
                auto p_in_tensor = std::dynamic_pointer_cast<phi::DenseTensor>(
                    in_tensor.impl());
                auto in_dense = *p_in_tensor;

                int64_t numel = in_dense.numel();
819 820
                int64_t send_numel = numel / nranks;
                int64_t offset = send_numel * rank_id;
821 822 823 824 825 826 827

                return self.AllGather(out_dense,
                                      in_dense,
                                      offset,
                                      send_numel,
                                      /*sync_op*/ true,
                                      /*use_calc_stream*/ true);
828 829
              },
              py::arg("out"),
830
              py::arg("in"),
831 832 833 834
              py::arg("num"),
              py::arg("id"),
              py::call_guard<py::gil_scoped_release>())

835
          .def(
L
LiYuRio 已提交
836
              "all_reduce_on_calc_stream",
837 838 839 840
              [](distributed::ProcessGroupStream &self,
                 py::handle py_tensor,
                 distributed::ReduceOp op) {
                auto tensor = CastPyArg2Tensor(py_tensor.ptr(), 0);
841
                auto p_dense =
842
                    std::dynamic_pointer_cast<phi::DenseTensor>(tensor.impl());
843 844 845 846 847
                auto in_dense = *p_dense;
                auto *out_dense = p_dense.get();
                distributed::AllreduceOptions opts{op};
                return self.AllReduce(out_dense,
                                      in_dense,
848 849 850 851 852
                                      opts,
                                      /*sync_op*/ true,
                                      /*use_calc_stream*/ true);
              },
              py::arg("tensor"),
L
LiYuRio 已提交
853
              py::arg("op") = distributed::ReduceOp::SUM,
854 855
              py::call_guard<py::gil_scoped_release>())

856
          .def(
L
LiYuRio 已提交
857
              "all_to_all_on_calc_stream",
858 859 860 861 862 863 864 865 866 867 868 869 870 871 872 873 874 875
              [](distributed::ProcessGroupStream &self,
                 py::handle py_in_tensor_list,
                 py::handle py_out_tensor_list) {
                auto in_tensor_list =
                    CastPyArg2VectorOfTensor(py_in_tensor_list.ptr(), 0);
                Tensor concat_in_tensor = paddle::concat(in_tensor_list, 0);
                auto in_dense = std::dynamic_pointer_cast<phi::DenseTensor>(
                    concat_in_tensor.impl());
                std::vector<phi::DenseTensor> in_wrapper = {*in_dense};

                auto out_tensor_list =
                    CastPyArg2VectorOfTensor(py_out_tensor_list.ptr(), 0);
                Tensor concat_out_tensor = paddle::concat(out_tensor_list, 0);
                auto out_dense = std::dynamic_pointer_cast<phi::DenseTensor>(
                    concat_out_tensor.impl());
                std::vector<phi::DenseTensor> out_wrapper = {*out_dense};

                // in_tensor_list must not be empty
876
                const auto &dev_ctx = self.GetDeviceContext(
877 878 879 880 881 882 883 884 885 886 887 888 889
                    in_tensor_list.back().place(), /*use_calc_stream*/ true);
                auto task = self.AllToAll(in_wrapper,
                                          out_wrapper,
                                          /*sync_op*/ true,
                                          /*use_calc_stream*/ true);
                distributed::SplitTensor(dev_ctx, *out_dense, &out_tensor_list);
                return task;
              },
              py::arg("in"),
              py::arg("out"),
              py::call_guard<py::gil_scoped_release>())

          .def(
L
LiYuRio 已提交
890
              "all_to_all_tensor_on_calc_stream",
891 892 893 894 895 896 897 898 899 900 901 902 903 904 905 906 907 908 909 910 911 912 913
              [](distributed::ProcessGroupStream &self,
                 py::handle py_in_tensor,
                 py::handle py_out_tensor) {
                auto in_tensor = CastPyArg2Tensor(py_in_tensor.ptr(), 0);
                auto in_dense = std::dynamic_pointer_cast<phi::DenseTensor>(
                    in_tensor.impl());
                std::vector<phi::DenseTensor> in_wrapper = {*in_dense};

                auto out_tensor = CastPyArg2Tensor(py_out_tensor.ptr(), 0);
                auto out_dense = std::dynamic_pointer_cast<phi::DenseTensor>(
                    out_tensor.impl());
                std::vector<phi::DenseTensor> out_wrapper = {*out_dense};

                return self.AllToAll(in_wrapper,
                                     out_wrapper,
                                     /*sync_op*/ true,
                                     /*use_calc_stream*/ true);
              },
              py::arg("in"),
              py::arg("out"),
              py::call_guard<py::gil_scoped_release>())

          .def(
L
LiYuRio 已提交
914
              "all_to_all_single_on_calc_stream",
915 916 917
              [](distributed::ProcessGroupStream &self,
                 py::handle py_in_tensor,
                 py::handle py_out_tensor,
918 919
                 std::vector<int64_t> &in_sizes,
                 std::vector<int64_t> &out_sizes) {
920
                auto in_tensor = CastPyArg2Tensor(py_in_tensor.ptr(), 0);
921 922 923 924
                auto in_dense = std::dynamic_pointer_cast<phi::DenseTensor>(
                    in_tensor.impl());
                std::vector<phi::DenseTensor> in_wrapper = {*in_dense};

925
                auto out_tensor = CastPyArg2Tensor(py_out_tensor.ptr(), 0);
926 927 928 929 930 931 932 933 934 935 936 937 938 939 940 941 942 943 944 945 946 947 948
                auto out_dense = std::dynamic_pointer_cast<phi::DenseTensor>(
                    out_tensor.impl());
                std::vector<phi::DenseTensor> out_wrapper = {*out_dense};

                return self.AllToAllSingle(in_wrapper,
                                           out_wrapper,
                                           in_sizes,
                                           out_sizes,
                                           /*sync_op*/ true,
                                           /*use_calc_stream*/ true);
              },
              py::arg("in"),
              py::arg("out"),
              py::arg("in_sizes"),
              py::arg("out_sizes"),
              py::call_guard<py::gil_scoped_release>())

          .def(
              "broadcast_on_calc_stream",
              [](distributed::ProcessGroupStream &self,
                 py::handle py_tensor,
                 int src) {
                auto tensor = CastPyArg2Tensor(py_tensor.ptr(), 0);
949
                auto p_dense =
950
                    std::dynamic_pointer_cast<phi::DenseTensor>(tensor.impl());
951 952 953 954 955
                auto *out_dense = p_dense.get();
                auto in_dense = *p_dense;
                distributed::BroadcastOptions opts{src};
                return self.Broadcast(out_dense,
                                      in_dense,
956 957 958 959 960 961 962 963 964 965 966 967 968 969 970 971 972 973 974 975 976 977 978 979 980 981 982 983 984 985 986 987 988 989 990 991 992 993 994 995 996 997 998 999 1000 1001 1002 1003 1004 1005 1006 1007 1008 1009 1010 1011 1012 1013 1014 1015 1016 1017 1018 1019 1020 1021 1022
                                      opts,
                                      /*sync_op*/ true,
                                      /*use_calc_stream*/ true);
              },
              py::arg("tensor"),
              py::arg("src"),
              py::call_guard<py::gil_scoped_release>())

          .def(
              "reduce_on_calc_stream",
              [](distributed::ProcessGroupStream &self,
                 py::handle py_in_tensor,
                 int dst,
                 distributed::ReduceOp op) {
                auto in_tensor = CastPyArg2Tensor(py_in_tensor.ptr(), 0);
                distributed::ReduceOptions opts{op, dst};
                auto dense = std::dynamic_pointer_cast<phi::DenseTensor>(
                    in_tensor.impl());
                std::vector<phi::DenseTensor> tensors = {*dense};
                return self.Reduce(tensors,
                                   tensors,
                                   opts,
                                   /*sync_op*/ true,
                                   /*use_calc_stream*/ true);
              },
              py::arg("tensor"),
              py::arg("dst"),
              py::arg("op"),
              py::call_guard<py::gil_scoped_release>())

          .def(
              "reduce_scatter_on_calc_stream",
              [](distributed::ProcessGroupStream &self,
                 py::handle py_in_tensor_list,
                 py::handle py_out_tensor,
                 distributed::ReduceOp op) {
                auto in_tensor_list =
                    CastPyArg2VectorOfTensor(py_in_tensor_list.ptr(), 0);
                Tensor concat_in_tensor = paddle::concat(in_tensor_list, 0);
                auto in_dense = std::dynamic_pointer_cast<phi::DenseTensor>(
                    concat_in_tensor.impl());
                std::vector<phi::DenseTensor> in_wrapper = {*in_dense};

                auto out_tensor = CastPyArg2Tensor(py_out_tensor.ptr(), 0);
                auto out_dense = std::dynamic_pointer_cast<phi::DenseTensor>(
                    out_tensor.impl());
                std::vector<phi::DenseTensor> out_wrapper = {*out_dense};

                distributed::ReduceScatterOptions opts{op};
                return self.ReduceScatter(in_wrapper,
                                          out_wrapper,
                                          opts,
                                          /*sync_op*/ true,
                                          /*use_calc_stream*/ true);
              },
              py::arg("in"),
              py::arg("out"),
              py::arg("op"),
              py::call_guard<py::gil_scoped_release>())

          .def(
              "reduce_scatter_tensor_on_calc_stream",
              [](distributed::ProcessGroupStream &self,
                 py::handle py_in_tensor,
                 py::handle py_out_tensor,
                 distributed::ReduceOp op) {
                auto in_tensor = CastPyArg2Tensor(py_in_tensor.ptr(), 0);
1023 1024
                auto in_dense = std::dynamic_pointer_cast<phi::DenseTensor>(
                    in_tensor.impl());
1025 1026 1027
                std::vector<phi::DenseTensor> in_wrapper = {*in_dense};

                auto out_tensor = CastPyArg2Tensor(py_out_tensor.ptr(), 0);
1028 1029
                auto out_dense = std::dynamic_pointer_cast<phi::DenseTensor>(
                    out_tensor.impl());
1030 1031 1032 1033 1034 1035 1036 1037
                std::vector<phi::DenseTensor> out_wrapper = {*out_dense};

                distributed::ReduceScatterOptions opts{op};
                return self.ReduceScatter(in_wrapper,
                                          out_wrapper,
                                          opts,
                                          /*sync_op*/ true,
                                          /*use_calc_stream*/ true);
1038 1039 1040
              },
              py::arg("in"),
              py::arg("out"),
1041 1042 1043 1044 1045 1046 1047 1048 1049 1050 1051 1052 1053 1054 1055 1056 1057 1058 1059 1060 1061 1062 1063 1064 1065 1066 1067 1068 1069 1070 1071 1072 1073 1074 1075 1076 1077 1078 1079 1080 1081 1082 1083 1084 1085 1086 1087 1088 1089 1090 1091 1092 1093 1094 1095 1096 1097 1098 1099
              py::arg("op"),
              py::call_guard<py::gil_scoped_release>())

          .def(
              "scatter_on_calc_stream",
              [](distributed::ProcessGroupStream &self,
                 py::handle py_in_tensor_list,
                 py::handle py_out_tensor,
                 int src) {
                auto in_tensor_list =
                    CastPyArg2VectorOfTensor(py_in_tensor_list.ptr(), 0);
                Tensor concat_in_tensor = paddle::concat(in_tensor_list, 0);
                auto in_dense = std::dynamic_pointer_cast<phi::DenseTensor>(
                    concat_in_tensor.impl());
                std::vector<phi::DenseTensor> in_wrapper = {*in_dense};

                auto out_tensor = CastPyArg2Tensor(py_out_tensor.ptr(), 0);
                auto out_dense = std::dynamic_pointer_cast<phi::DenseTensor>(
                    out_tensor.impl());
                std::vector<phi::DenseTensor> out_wrapper = {*out_dense};

                distributed::ScatterOptions opts{src};
                return self.Scatter(in_wrapper,
                                    out_wrapper,
                                    opts,
                                    /*sync_op*/ true,
                                    /*use_calc_stream*/ true);
              },
              py::arg("in"),
              py::arg("out"),
              py::arg("src"),
              py::call_guard<py::gil_scoped_release>())

          .def(
              "scatter_tensor_on_calc_stream",
              [](distributed::ProcessGroupStream &self,
                 py::handle py_in_tensor,
                 py::handle py_out_tensor,
                 int src) {
                auto in_tensor = CastPyArg2Tensor(py_in_tensor.ptr(), 0);
                auto in_dense = std::dynamic_pointer_cast<phi::DenseTensor>(
                    in_tensor.impl());
                std::vector<phi::DenseTensor> in_wrapper = {*in_dense};

                auto out_tensor = CastPyArg2Tensor(py_out_tensor.ptr(), 0);
                auto out_dense = std::dynamic_pointer_cast<phi::DenseTensor>(
                    out_tensor.impl());
                std::vector<phi::DenseTensor> out_wrapper = {*out_dense};

                distributed::ScatterOptions opts{src};
                return self.Scatter(in_wrapper,
                                    out_wrapper,
                                    opts,
                                    /*sync_op*/ true,
                                    /*use_calc_stream*/ true);
              },
              py::arg("in"),
              py::arg("out"),
              py::arg("src"),
1100 1101
              py::call_guard<py::gil_scoped_release>())

1102 1103 1104 1105 1106 1107
          .def(
              "send_on_calc_stream",
              [](distributed::ProcessGroupStream &self,
                 py::handle py_tensor,
                 int dst) {
                auto tensor = CastPyArg2Tensor(py_tensor.ptr(), 0);
1108
                auto p_dense =
1109
                    std::dynamic_pointer_cast<phi::DenseTensor>(tensor.impl());
1110
                auto *out_dense = p_dense.get();
1111
                // numel == -1 indicates sending the whole tensor
1112
                return self.Send(out_dense,
1113
                                 dst,
1114 1115
                                 /*offset*/ 0,
                                 /*numel*/ -1,
1116 1117 1118 1119 1120 1121 1122 1123 1124 1125 1126 1127 1128 1129 1130
                                 /*sync_op*/ true,
                                 /*use_calc_stream*/ true);
              },
              py::arg("tensor"),
              py::arg("dst"),
              py::call_guard<py::gil_scoped_release>())

          .def(
              "send_partial_on_calc_stream",
              [](distributed::ProcessGroupStream &self,
                 py::handle py_tensor,
                 int dst_rank,
                 int nranks,
                 int rank_id) {
                auto tensor = CastPyArg2Tensor(py_tensor.ptr(), 0);
1131
                auto p_dense =
1132
                    std::dynamic_pointer_cast<phi::DenseTensor>(tensor.impl());
1133 1134
                auto *out_dense = p_dense.get();

1135
                int64_t numel = p_dense->numel();
1136 1137
                int64_t send_numel = numel / nranks;
                int64_t offset = send_numel * rank_id;
1138 1139 1140 1141 1142 1143 1144

                return self.Send(out_dense,
                                 dst_rank,
                                 offset,
                                 send_numel,
                                 /*sync_op*/ true,
                                 /*use_calc_stream*/ true);
1145 1146 1147 1148 1149 1150 1151 1152 1153 1154 1155 1156 1157
              },
              py::arg("tensor"),
              py::arg("dst"),
              py::arg("num"),
              py::arg("id"),
              py::call_guard<py::gil_scoped_release>())

          .def(
              "recv_on_calc_stream",
              [](distributed::ProcessGroupStream &self,
                 py::handle py_tensor,
                 int src) {
                auto tensor = CastPyArg2Tensor(py_tensor.ptr(), 0);
1158
                auto p_dense =
1159
                    std::dynamic_pointer_cast<phi::DenseTensor>(tensor.impl());
1160
                auto *in_dense = p_dense.get();
1161
                // numel == -1 indicates receiving the whole tensor
1162
                return self.Recv(in_dense,
1163
                                 src,
1164 1165
                                 /*offset*/ 0,
                                 /*numel*/ -1,
1166 1167 1168 1169 1170 1171 1172 1173 1174 1175 1176 1177 1178 1179 1180
                                 /*sync_op*/ true,
                                 /*use_calc_stream*/ true);
              },
              py::arg("tensor"),
              py::arg("src"),
              py::call_guard<py::gil_scoped_release>())

          .def(
              "recv_partial_on_calc_stream",
              [](distributed::ProcessGroupStream &self,
                 py::handle py_tensor,
                 int src_rank,
                 int nranks,
                 int rank_id) {
                auto tensor = CastPyArg2Tensor(py_tensor.ptr(), 0);
1181
                auto p_dense =
1182
                    std::dynamic_pointer_cast<phi::DenseTensor>(tensor.impl());
1183 1184
                auto *out_dense = p_dense.get();

1185
                int64_t numel = p_dense->numel();
1186 1187
                int64_t recv_numel = numel / nranks;
                int64_t offset = recv_numel * rank_id;
1188 1189 1190 1191 1192 1193 1194

                return self.Recv(out_dense,
                                 src_rank,
                                 offset,
                                 recv_numel,
                                 /*sync_op*/ true,
                                 /*use_calc_stream*/ true);
1195 1196 1197 1198 1199
              },
              py::arg("tensor"),
              py::arg("src"),
              py::arg("num"),
              py::arg("id"),
1200 1201
              py::call_guard<py::gil_scoped_release>());

1202
#if defined(PADDLE_WITH_RCCL) || defined(PADDLE_WITH_NCCL)
1203 1204 1205
  auto processGroupNCCL =
      py::class_<distributed::ProcessGroupNCCL,
                 std::shared_ptr<distributed::ProcessGroupNCCL>>(
1206
          *m, "ProcessGroupNCCL", ProcessGroupStream)
1207 1208 1209 1210 1211 1212 1213 1214 1215 1216 1217 1218 1219 1220 1221
          .def(py::init<const std::shared_ptr<distributed::Store> &,
                        int,
                        int,
                        int>(),
               py::arg("store"),
               py::arg("rank"),
               py::arg("world_size"),
               py::arg("group_id") = 0,
               py::call_guard<py::gil_scoped_release>());

  processGroupNCCL.def_static(
      "group_start", []() { distributed::ProcessGroupNCCL::GroupStart(); });
  processGroupNCCL.def_static(
      "group_end", []() { distributed::ProcessGroupNCCL::GroupEnd(); });

1222
#endif
1223

W
wuhuachaocoding 已提交
1224 1225 1226 1227 1228 1229 1230 1231 1232 1233 1234 1235 1236 1237 1238 1239 1240 1241 1242
#if defined(PADDLE_WITH_MPI)
  py::class_<distributed::ProcessGroupMPI,
             std::shared_ptr<distributed::ProcessGroupMPI>>(
      *m, "ProcessGroupMPI", ProcessGroup)
      .def_static(
          "create",
          [](const std::vector<int> &ranks,
             int gid) -> std::shared_ptr<distributed::ProcessGroupMPI> {
            return paddle::distributed::ProcessGroupMPI::CreateProcessGroupMPI(
                ranks, gid);
          })
      .def("get_rank",
           &distributed::ProcessGroup::GetRank,
           py::call_guard<py::gil_scoped_release>())
      .def("get_world_size",
           &distributed::ProcessGroup::GetSize,
           py::call_guard<py::gil_scoped_release>());
#endif

1243 1244 1245 1246 1247
#if defined(PADDLE_WITH_CUSTOM_DEVICE)
  py::class_<distributed::ProcessGroupCustom,
             std::shared_ptr<distributed::ProcessGroupCustom>>(
      *m, "ProcessGroupCustom", ProcessGroup)
      .def(py::init<const std::shared_ptr<distributed::Store> &,
1248
                    const std::string &,
1249 1250 1251 1252
                    int,
                    int,
                    int>(),
           py::arg("store"),
1253
           py::arg("device_type"),
1254 1255 1256 1257 1258
           py::arg("rank"),
           py::arg("world_size"),
           py::arg("group_id") = 0,
           py::call_guard<py::gil_scoped_release>());

1259 1260
#endif

J
james 已提交
1261 1262 1263 1264 1265 1266 1267 1268 1269 1270 1271 1272 1273 1274 1275 1276
#if defined(PADDLE_WITH_XPU_BKCL)
  auto processGroupBKCL =
      py::class_<distributed::ProcessGroupBKCL,
                 std::shared_ptr<distributed::ProcessGroupBKCL>>(
          *m, "ProcessGroupBKCL", ProcessGroup)
          .def(py::init<const std::shared_ptr<distributed::Store> &,
                        int,
                        int,
                        int>(),
               py::arg("store"),
               py::arg("rank"),
               py::arg("world_size"),
               py::arg("group_id") = 0,
               py::call_guard<py::gil_scoped_release>());
#endif

1277 1278 1279
  py::class_<distributed::ProcessGroup::Task,
             std::shared_ptr<distributed::ProcessGroup::Task>>(*m, "task")
      .def("is_completed", &distributed::ProcessGroup::Task::IsCompleted)
1280
      .def("is_sync", &distributed::ProcessGroup::Task::IsSync)
1281 1282
      .def("wait",
           &distributed::ProcessGroup::Task::Wait,
1283 1284
           py::arg("timeout") = kWaitTimeout,
           py::call_guard<py::gil_scoped_release>())
1285 1286
      .def("synchronize",
           &distributed::ProcessGroup::Task::Synchronize,
1287 1288
           py::call_guard<py::gil_scoped_release>());

1289 1290 1291
#if defined(PADDLE_WITH_GLOO)
  py::class_<ProcessGroupGloo, std::shared_ptr<ProcessGroupGloo>>(
      *m, "ProcessGroupGloo", ProcessGroup)
1292 1293 1294 1295
      .def(py::init<const std::shared_ptr<paddle::distributed::Store> &,
                    int,
                    int,
                    int,
1296
                    std::shared_ptr<GlooOptions> &>(),
1297
           py::call_guard<py::gil_scoped_release>())
1298
      .def(py::init([](const std::shared_ptr<paddle::distributed::Store> &store,
1299 1300 1301
                       int rank,
                       int world_size,
                       int gid) {
1302 1303 1304 1305 1306 1307 1308 1309
             auto opts = GlooOptions::create();
             char *ifname = getenv(GLOO_SOCKET_IFNAME_ENV.c_str());
             if (ifname && strlen(ifname) > 1) {
               opts->device = ProcessGroupGloo::createDeviceForInterface(
                   std::string(ifname));
             } else {
               opts->device = ProcessGroupGloo::createDefaultDevice();
             }
1310
             return std::make_shared<ProcessGroupGloo>(
1311
                 store, rank, world_size, gid, opts);
1312
           }),
1313 1314 1315 1316
           py::arg("store"),
           py::arg("rank"),
           py::arg("world_size"),
           py::arg("group_id") = 0,
1317
           py::call_guard<py::gil_scoped_release>())
1318 1319 1320 1321
      .def_static("create_default_device",
                  &ProcessGroupGloo::createDefaultDevice);
#endif

1322 1323
  m->def(
      "eager_assign_group_by_size",
1324 1325
      [](py::handle py_tensors,
         std::vector<bool> is_sparse_gradient,
1326 1327 1328 1329 1330 1331
         std::vector<size_t> group_size_limits,
         std::vector<int64_t> tensor_indices) {
        auto tensors = CastPyArg2VectorOfTensor(py_tensors.ptr(), 0);
        return distributed::Eager_AssignGroupBySize(
            tensors, is_sparse_gradient, group_size_limits, tensor_indices);
      },
1332 1333
      py::arg("tensors"),
      py::arg("is_sparse_gradient"),
1334 1335 1336
      py::arg("group_size_limits") = std::vector<size_t>{25 * 1024 * 1024},
      py::arg("tensor_indices") = std::vector<int64_t>{},
      py::call_guard<py::gil_scoped_release>());
1337 1338

  py::class_<distributed::EagerReducer,
1339 1340
             std::shared_ptr<distributed::EagerReducer>>(
      *m, "EagerReducer", R"DOC()DOC")
1341
      .def(py::init(&CreateEagerReducer))
1342 1343
      .def(
          "prepare_for_backward",
1344
          [](distributed::EagerReducer &self, py::handle py_tensors) {
1345
            auto params = CastPyArg2VectorOfTensor(py_tensors.ptr(), 0);
1346
            self.PrepareForBackward(params);
1347
          },
1348 1349
          py::arg("tensors"),
          py::call_guard<py::gil_scoped_release>());
1350 1351 1352 1353
}

}  // end namespace pybind
}  // namespace paddle