未验证 提交 a0dffd39 编写于 作者: L LiYuRio 提交者: GitHub

Move group and all reduce from collective to communication (#45848)

上级 45b93325
......@@ -293,6 +293,14 @@ std::shared_ptr<ProcessGroup::Task> ProcessGroupGloo::AllReduce(
std::vector<phi::DenseTensor>& inputs,
std::vector<phi::DenseTensor>& outputs,
const AllreduceOptions& opts) {
return AllReduce(inputs, outputs, opts, true);
}
std::shared_ptr<ProcessGroup::Task> ProcessGroupGloo::AllReduce(
std::vector<phi::DenseTensor>& inputs,
std::vector<phi::DenseTensor>& outputs,
const AllreduceOptions& opts,
bool sync_op) {
auto tag = next_tag();
std::shared_ptr<GlooTask> task;
auto context = get_context();
......
......@@ -120,6 +120,12 @@ class ProcessGroupGloo : public ProcessGroup {
std::vector<phi::DenseTensor>& outputs,
const AllreduceOptions& opts = AllreduceOptions()) override;
std::shared_ptr<ProcessGroup::Task> AllReduce(
std::vector<phi::DenseTensor>& inputs,
std::vector<phi::DenseTensor>& outputs,
const AllreduceOptions& opts,
bool sync_op) override;
std::shared_ptr<ProcessGroup::Task> Barrier(
const BarrierOptions& = BarrierOptions()) override;
......
......@@ -52,54 +52,12 @@ from .fleet.layers.mpu.mp_ops import _c_softmax_with_cross_entropy
from .fleet.layers.mpu.mp_ops import _linear
from .fleet.layers.mpu.mp_ops import _parallel_linear
from .fleet.layers.mpu.mp_ops import _parallel_embedding
from .communication.comm_utils import ReduceOp
from .communication.group import Group, _add_new_group
from .communication.all_reduce import all_reduce
from .communication.reduce import _get_reduce_op, ReduceOp
__all__ = []
class Group():
"""
The abstract representation of group.
"""
def __init__(self, rank, rank_num, id=0, ranks=[], pg=None, name=None):
self.rank = rank
self.nranks = rank_num
self.id = id
self.ranks = ranks
self.pg = pg
self.name = name
def is_member(self):
if self.rank < 0:
return False
if self.nranks < 2:
return False
return True
def get_group_rank(self, rank):
if self.is_member() and rank in self.ranks:
return self.ranks.index(rank)
else:
return -1
@property
def process_group(self):
return self.pg
@property
def world_size(self):
return self.nranks if self.rank >= 0 else -1
def __repr__(self):
debug_str = "rank: {}, nranks: {}, id: {}, ranks: ".format(
self.rank, self.nranks, self.id)
debug_str += ", ".join(map(str, self.ranks))
debug_str += "; name: "
debug_str += self.name if self.name else "None"
return debug_str
_global_env = None
......@@ -147,9 +105,8 @@ def _get_group_map():
global _group_map
if _global_env_gid not in _group_map:
genv = _get_global_env()
_group_map[_global_env_gid] = Group(genv.rank,
genv.world_size,
ranks=list(range(genv.world_size)))
_group_map[_global_env_gid] = Group(genv.rank, 0,
list(range(genv.world_size)))
return _group_map
......@@ -197,19 +154,6 @@ def _new_ring_id():
return len(_get_group_map()) + max(_get_global_env().nrings, 9)
def _get_reduce_op(reduce_op, func_name):
if reduce_op == ReduceOp.SUM:
return core.ReduceOp.SUM
elif reduce_op == ReduceOp.MAX:
return core.ReduceOp.MAX
elif reduce_op == ReduceOp.MIN:
return core.ReduceOp.MIN
elif reduce_op == ReduceOp.PROD:
return core.ReduceOp.PRODUCT
else:
raise ValueError("Unknown reduce_op type for {}.".format(func_name))
def get_group(id=0):
"""
......@@ -451,10 +395,13 @@ def new_group(ranks=None, backend=None, timeout=_default_timeout):
else:
rank = -1
pg = None
group = Group(rank, size, id=gid, ranks=ranks, pg=pg, name=group_name)
group = Group(rank, gid, ranks, pg=pg, name=group_name)
_group_map_by_name[group_name] = group
_group_map[gid] = group
_group_map_backend[group] = backend
#TODO: The method below is a new method for group management, will replace the previous
# three in the future.
_add_new_group(group)
# TODO(shenliang03): This is a temporary solution to solve the problem of
# hang caused by tcp
......@@ -476,13 +423,13 @@ def new_group(ranks=None, backend=None, timeout=_default_timeout):
ring_id = _new_ring_id()
if global_rank not in ranks:
gp = Group(-1, -1, ring_id, ranks)
gp = Group(-1, ring_id, ranks)
_group_map[ring_id] = gp
else:
ranks = sorted(ranks)
group_rank = ranks.index(global_rank)
group_size = len(ranks)
gp = Group(group_rank, group_size, ring_id, ranks)
gp = Group(group_rank, ring_id, ranks)
_group_map[ring_id] = gp
if group_size >= 2:
......@@ -748,104 +695,6 @@ def broadcast(tensor, src, group=None, sync_op=True):
})
def all_reduce(tensor, op=ReduceOp.SUM, group=None, sync_op=True):
"""
Reduce a tensor over all ranks so that all get the result.
As shown below, one process is started with a GPU and the data of this process is represented
by its group rank. The reduce operator is sum. Through all_reduce operator,
each GPU will have the sum of the data from all GPUs.
.. image:: https://githubraw.cdn.bcebos.com/PaddlePaddle/docs/develop/docs/api/paddle/distributed/img/allreduce.png
:width: 800
:alt: all_reduce
:align: center
Args:
tensor (Tensor): The input Tensor. It also works as the output Tensor. Its data type
should be float16, float32, float64, int32, int64, int8, uint8 or bool.
op (ReduceOp.SUM|ReduceOp.MAX|ReduceOp.MIN|ReduceOp.PROD, optional): The operation used. Default value is ReduceOp.SUM.
group (Group, optional): The group instance return by new_group or None for global default group.
sync_op (bool, optional): Wether this op is a sync op. Default value is True.
Returns:
None.
Examples:
.. code-block:: python
# required: distributed
import paddle
import paddle.distributed as dist
dist.init_parallel_env()
if dist.get_rank() == 0:
data = paddle.to_tensor([[4, 5, 6], [4, 5, 6]])
else:
data = paddle.to_tensor([[1, 2, 3], [1, 2, 3]])
dist.all_reduce(data)
print(data)
# [[5, 7, 9], [5, 7, 9]] (2 GPUs)
"""
if group is not None and not group.is_member():
return
if in_dygraph_mode():
op_type = _get_reduce_op(op, "all_reduce")
group = _get_default_group() if group is None else group
task = group.process_group.allreduce(tensor, op_type)
if sync_op:
task.wait()
return None
else:
return task
use_calc_stream = sync_op
ring_id = 0 if group is None else group.id
if _non_static_mode():
if op == ReduceOp.SUM:
return _legacy_C_ops.c_allreduce_sum_(tensor, 'use_calc_stream',
use_calc_stream, 'ring_id',
ring_id)
elif op == ReduceOp.MAX:
return _legacy_C_ops.c_allreduce_max_(tensor, 'use_calc_stream',
use_calc_stream, 'ring_id',
ring_id)
elif op == ReduceOp.MIN:
return _legacy_C_ops.c_allreduce_min_(tensor, 'use_calc_stream',
use_calc_stream, 'ring_id',
ring_id)
elif op == ReduceOp.PROD:
return _legacy_C_ops.c_allreduce_prod_(tensor, 'use_calc_stream',
use_calc_stream, 'ring_id',
ring_id)
else:
raise ValueError("Unknown parameter: {}.".format(op))
check_variable_and_dtype(tensor, 'tensor', [
'float16', 'float32', 'float64', 'int32', 'int64', 'int8', 'uint8',
'bool'
], 'all_reduce')
if op == ReduceOp.SUM:
op_type = 'c_allreduce_sum'
elif op == ReduceOp.MAX:
op_type = 'c_allreduce_max'
elif op == ReduceOp.MIN:
op_type = 'c_allreduce_min'
elif op == ReduceOp.PROD:
op_type = 'c_allreduce_prod'
if not isinstance(ring_id, int):
raise ValueError("The type of 'ring_id' for all_reduce should be int.")
helper = LayerHelper(op_type, **locals())
helper.append_op(type=op_type,
inputs={'X': [tensor]},
outputs={'Out': [tensor]},
attrs={
'ring_id': ring_id,
'use_calc_stream': use_calc_stream
})
def reduce(tensor, dst, op=ReduceOp.SUM, group=None, sync_op=True):
"""
......
# 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.
import paddle
import paddle.fluid.framework as framework
from paddle.distributed.communication import stream as stream
from paddle.distributed.communication.reduce import ReduceOp
def all_reduce(tensor, op=ReduceOp.SUM, group=None, sync_op=True):
"""
Reduce a tensor over all ranks so that all get the result.
As shown below, one process is started with a GPU and the data of this process is represented
by its group rank. The reduce operator is sum. Through all_reduce operator,
each GPU will have the sum of the data from all GPUs.
.. image:: https://githubraw.cdn.bcebos.com/PaddlePaddle/docs/develop/docs/api/paddle/distributed/img/allreduce.png
:width: 800
:alt: all_reduce
:align: center
Args:
tensor (Tensor): The input Tensor. It also works as the output Tensor. Its data type
should be float16, float32, float64, int32, int64, int8, uint8 or bool.
op (ReduceOp.SUM|ReduceOp.MAX|ReduceOp.MIN|ReduceOp.PROD, optional): The operation used. Default value is ReduceOp.SUM.
group (Group, optional): The group instance return by new_group or None for global default group.
sync_op (bool, optional): Wether this op is a sync op. Default value is True.
Returns:
Return a task object.
Examples:
.. code-block:: python
# required: distributed
import paddle
import paddle.distributed as dist
dist.init_parallel_env()
if dist.get_rank() == 0:
data = paddle.to_tensor([[4, 5, 6], [4, 5, 6]])
else:
data = paddle.to_tensor([[1, 2, 3], [1, 2, 3]])
dist.all_reduce(data)
print(data)
# [[5, 7, 9], [5, 7, 9]] (2 GPUs)
"""
if not framework._in_legacy_dygraph():
return stream.all_reduce(tensor,
op=op,
group=group,
sync_op=sync_op,
use_calc_stream=False)
# code below will be removed after we remove the old dygraph
use_calc_stream = sync_op
ring_id = 0 if group is None else group.id
if op == ReduceOp.SUM:
return paddle._legacy_C_ops.c_allreduce_sum_(tensor, 'use_calc_stream',
use_calc_stream, 'ring_id',
ring_id)
elif op == ReduceOp.MAX:
return paddle._legacy_C_ops.c_allreduce_max_(tensor, 'use_calc_stream',
use_calc_stream, 'ring_id',
ring_id)
elif op == ReduceOp.MIN:
return paddle._legacy_C_ops.c_allreduce_min_(tensor, 'use_calc_stream',
use_calc_stream, 'ring_id',
ring_id)
elif op == ReduceOp.PROD:
return paddle._legacy_C_ops.c_allreduce_prod_(tensor, 'use_calc_stream',
use_calc_stream,
'ring_id', ring_id)
else:
raise ValueError("Unknown parameter: {}.".format(op))
# 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.
class Group():
"""
The abstract representation of group.
"""
def __init__(self, rank_in_group, id, ranks, pg=None, name=None):
self._rank_in_group = rank_in_group
self._world_size = len(ranks) if rank_in_group >= 0 else -1
self._id = id
self._ranks = ranks
self._pg = pg
self._name = name
@property
def rank(self):
return self._rank_in_group
@property
def ranks(self):
return self._ranks
@property
def nranks(self):
return len(self._ranks)
@property
def name(self):
return self._name
@property
def process_group(self):
return self._pg
@property
def world_size(self):
return self._world_size
@property
def id(self):
return self._id
def is_member(self):
if self.rank < 0:
return False
if self.nranks < 2:
return False
return True
def get_group_rank(self, rank):
if self.is_member():
return self.ranks.index(rank)
else:
return -1
def __repr__(self):
debug_str = "rank: {}, nranks: {}, id: {}, ranks: ".format(
self.rank, self.nranks, self.id)
debug_str += ", ".join(map(str, self.ranks))
debug_str += "; name: "
debug_str += self.name if self.name else "None"
return debug_str
class _GroupManager():
global_group_id = 0
group_map_by_id = {}
def _get_global_group():
if _GroupManager.global_group_id not in _GroupManager.group_map_by_id:
raise RuntimeError("The global group is not initialized.")
return _GroupManager.group_map_by_id[_GroupManager.global_group_id]
def _add_new_group(group):
if group.id in _GroupManager.group_map_by_id:
raise RuntimeError("The group with id {} already exist.".format(
group.id))
_GroupManager.group_map_by_id[group.id] = group
# Copyright (c) 2022 PaddlePaddle Authors. All Rights Reserved.
# 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.
......@@ -12,6 +12,9 @@
# See the License for the specific language governing permissions and
# limitations under the License.
import paddle.fluid.framework as framework
import paddle.fluid.core as core
class ReduceOp:
"""
......@@ -48,3 +51,26 @@ class ReduceOp:
MIN = 2
PROD = 3
AVG = 4
def _get_reduce_op(reduce_op, func_name):
if framework.in_dygraph_mode():
if reduce_op == ReduceOp.SUM:
return core.ReduceOp.SUM
elif reduce_op == ReduceOp.MAX:
return core.ReduceOp.MAX
elif reduce_op == ReduceOp.MIN:
return core.ReduceOp.MIN
elif reduce_op == ReduceOp.PROD:
return core.ReduceOp.PRODUCT
else:
if reduce_op == ReduceOp.SUM:
return 'c_allreduce_sum'
elif reduce_op == ReduceOp.MAX:
return 'c_allreduce_max'
elif reduce_op == ReduceOp.MIN:
return 'c_allreduce_min'
elif reduce_op == ReduceOp.PROD:
return 'c_allreduce_prod'
raise ValueError("Unknown reduce_op type for {}.".format(func_name))
......@@ -13,12 +13,16 @@
# limitations under the License.
import paddle.fluid.framework as framework
from paddle.distributed import collective
import paddle.fluid.data_feeder as data_feeder
import paddle.fluid.layer_helper as layer_helper
from paddle.distributed.communication.reduce import _get_reduce_op, ReduceOp
from paddle.distributed.communication.group import _get_global_group
def _all_reduce_in_dygraph(tensor, op, group, sync_op, use_calc_stream):
op_type = collective._get_reduce_op(op, "all_reduce")
group = collective._get_default_group() if group is None else group
op_type = _get_reduce_op(op, "all_reduce")
group = _get_global_group() if group is None else group
if use_calc_stream:
return group.process_group.allreduce_on_calc_stream(tensor, op_type)
......@@ -29,8 +33,34 @@ def _all_reduce_in_dygraph(tensor, op, group, sync_op, use_calc_stream):
return task
def _all_reduce_in_static_mode(tensor, op, group, sync_op, use_calc_stream):
data_feeder.check_variable_and_dtype(tensor, 'tensor', [
'float16', 'float32', 'float64', 'int32', 'int64', 'int8', 'uint8',
'bool'
], 'all_reduce')
op_type = _get_reduce_op(op, "all_reduce")
ring_id = 0 if group is None else group.id
if not isinstance(ring_id, int):
raise ValueError("The type of 'ring_id' for all_reduce should be int.")
# TODO: Support task and use task.wait in static mode
# Use use_calc_stream rather than sync_op
helper = layer_helper.LayerHelper(op_type, **locals())
helper.append_op(type=op_type,
inputs={'X': [tensor]},
outputs={'Out': [tensor]},
attrs={
'ring_id': ring_id,
'use_calc_stream': sync_op
})
return None
def all_reduce(tensor,
op=collective.ReduceOp.SUM,
op=ReduceOp.SUM,
group=None,
sync_op=True,
use_calc_stream=False):
......@@ -41,7 +71,7 @@ def all_reduce(tensor,
Args:
tensor (Tensor): The input tensor on each rank. The result will overwrite this tenor after communication. Support
float16, float32, float64, int32 or int64 as the input data type.
op (ReduceOp.SUM|ReduceOp.MAX|ReduceOp.Min|ReduceOp.PROD, optional): The reduction used. If none is given, use ReduceOp.SUM as default.
op (ReduceOp.SUM|ReduceOp.MAX|ReduceOp.MIN|ReduceOp.PROD, optional): The reduction used. If none is given, use ReduceOp.SUM as default.
group (Group, optional): Communicate in which group. If none is given, use the global group as default.
sync_op (bool, optional): Indicate whether the communication is sync or not. If none is given, use true as default.
use_calc_stream (bool, optional): Indicate whether the communication is done on calculation stream. If none is given, use false as default. This
......@@ -50,9 +80,6 @@ def all_reduce(tensor,
Returns:
Return a task object.
Warning:
This API only supports the dygraph mode now.
Examples:
.. code-block:: python
......@@ -84,7 +111,6 @@ def all_reduce(tensor,
if framework.in_dygraph_mode():
return _all_reduce_in_dygraph(tensor, op, group, sync_op,
use_calc_stream)
raise RuntimeError(
"paddle.distributed.stream.all_reduce is only supported in dygraph mode now."
)
else:
return _all_reduce_in_static_mode(tensor, op, group, sync_op,
use_calc_stream)
......@@ -377,8 +377,8 @@ class _CommunicateGroup(object):
def set_comm_group(self, group_name, group_rank, group_size, ring_id,
group_ranks):
group = paddle.distributed.collective.Group(group_rank, group_size,
ring_id, group_ranks)
group = paddle.distributed.collective.Group(group_rank, ring_id,
group_ranks)
self.groups[group_name] = group
def get_group(self, group_name):
......
......@@ -22,7 +22,7 @@ from paddle.fluid.layer_helper import LayerHelper
from paddle.fluid.data_feeder import check_variable_and_dtype
from paddle.fluid.dygraph import layers
from paddle.distributed import collective
from ....communication.comm_utils import ReduceOp
from ....communication.reduce import ReduceOp
from paddle.fluid.data_feeder import check_dtype
import paddle.fluid.dygraph_utils as dygraph_utils
......
......@@ -43,6 +43,7 @@ from paddle.distributed.collective import _set_default_store
from paddle.distributed.collective import _new_process_group_impl
from paddle.distributed.collective import Group
from paddle.distributed.collective import _set_group_map_backend
from paddle.distributed.communication.group import _add_new_group
__all__ = []
......@@ -258,15 +259,11 @@ def init_parallel_env():
_default_group_name,
pg_options=None)
ranks = list(range(world_size))
group = Group(rank,
world_size,
id=0,
ranks=ranks,
pg=pg,
name=_default_group_name)
group = Group(rank, 0, ranks, pg=pg, name=_default_group_name)
_set_group_map_by_name(_default_group_name, group)
_set_group_map(0, group)
_set_group_map_backend(group, backend)
_add_new_group(group)
parallel_helper._set_parallel_ctx(True)
paddle.distributed.barrier(group=group)
......
......@@ -265,7 +265,6 @@ class MoELayer(nn.Layer):
from paddle.distributed import fleet
moe_group = Group(fleet.worker_index(),
fleet.worker_num(),
0,
list(range(fleet.worker_num())))
mp_group = None
......
Markdown is supported
0% .
You are about to add 0 people to the discussion. Proceed with caution.
先完成此消息的编辑!
想要评论请 注册