Skip to content
体验新版
项目
组织
正在加载...
登录
切换导航
打开侧边栏
PaddlePaddle
Paddle
提交
51ba2a0f
P
Paddle
项目概览
PaddlePaddle
/
Paddle
大约 1 年 前同步成功
通知
2299
Star
20931
Fork
5422
代码
文件
提交
分支
Tags
贡献者
分支图
Diff
Issue
1423
列表
看板
标记
里程碑
合并请求
543
Wiki
0
Wiki
分析
仓库
DevOps
项目成员
Pages
P
Paddle
项目概览
项目概览
详情
发布
仓库
仓库
文件
提交
分支
标签
贡献者
分支图
比较
Issue
1,423
Issue
1,423
列表
看板
标记
里程碑
合并请求
543
合并请求
543
Pages
分析
分析
仓库分析
DevOps
Wiki
0
Wiki
成员
成员
收起侧边栏
关闭侧边栏
动态
分支图
创建新Issue
提交
Issue看板
未验证
提交
51ba2a0f
编写于
8月 31, 2023
作者:
C
caozhou
提交者:
GitHub
8月 31, 2023
浏览文件
操作
浏览文件
下载
电子邮件补丁
差异文件
add op cost interface (#56803)
上级
d53972fd
变更
5
显示空白变更内容
内联
并排
Showing
5 changed file
with
250 addition
and
4 deletion
+250
-4
python/paddle/distributed/auto_parallel/static/cluster.py
python/paddle/distributed/auto_parallel/static/cluster.py
+2
-0
python/paddle/distributed/auto_parallel/static/cost/__init__.py
.../paddle/distributed/auto_parallel/static/cost/__init__.py
+1
-0
python/paddle/distributed/auto_parallel/static/cost/base_cost.py
...paddle/distributed/auto_parallel/static/cost/base_cost.py
+61
-4
test/auto_parallel/CMakeLists.txt
test/auto_parallel/CMakeLists.txt
+1
-0
test/auto_parallel/test_cost_interface.py
test/auto_parallel/test_cost_interface.py
+185
-0
未找到文件。
python/paddle/distributed/auto_parallel/static/cluster.py
浏览文件 @
51ba2a0f
...
@@ -429,6 +429,7 @@ class Cluster:
...
@@ -429,6 +429,7 @@ class Cluster:
# This property only be valid when the cluster consists of machines,
# This property only be valid when the cluster consists of machines,
# which have the same number accelerators.
# which have the same number accelerators.
self
.
_num_devices_per_machine
=
None
self
.
_num_devices_per_machine
=
None
self
.
_gpu_model
=
None
def
gen_default_config_cluster
(
def
gen_default_config_cluster
(
self
,
self
,
...
@@ -451,6 +452,7 @@ class Cluster:
...
@@ -451,6 +452,7 @@ class Cluster:
dcu_models
=
[
"DCU"
]
dcu_models
=
[
"DCU"
]
all_gpu_models
=
gpu_models
+
xpu_models
+
dcu_models
all_gpu_models
=
gpu_models
+
xpu_models
+
dcu_models
self
.
_num_devices_per_machine
=
device_count
self
.
_num_devices_per_machine
=
device_count
self
.
_gpu_model
=
gpu_model
def
_convert_to_type
(
gpu_model
):
def
_convert_to_type
(
gpu_model
):
type
=
None
type
=
None
...
...
python/paddle/distributed/auto_parallel/static/cost/__init__.py
浏览文件 @
51ba2a0f
...
@@ -22,6 +22,7 @@ from .base_cost import build_comp_desc_from_dist_op
...
@@ -22,6 +22,7 @@ from .base_cost import build_comp_desc_from_dist_op
from
.base_cost
import
build_comm_desc_from_dist_op
from
.base_cost
import
build_comm_desc_from_dist_op
from
.base_cost
import
build_comm_costs_from_descs
from
.base_cost
import
build_comm_costs_from_descs
from
.base_cost
import
build_comp_costs_from_descs
from
.base_cost
import
build_comp_costs_from_descs
from
.base_cost
import
calc_time_by_cost_model
from
.comp_op_cost
import
EmbeddingOpCost
from
.comp_op_cost
import
EmbeddingOpCost
from
.comp_op_cost
import
EmbeddingGradOpCost
from
.comp_op_cost
import
EmbeddingGradOpCost
...
...
python/paddle/distributed/auto_parallel/static/cost/base_cost.py
浏览文件 @
51ba2a0f
...
@@ -19,7 +19,7 @@ import numpy as np
...
@@ -19,7 +19,7 @@ import numpy as np
import
paddle
import
paddle
from
paddle.utils.flops
import
flops
from
paddle.utils.flops
import
flops
from
..cluster
import
LinkType
from
..cluster
import
LinkType
,
get_default_cluster
from
..dist_tensor
import
DistributedTensor
from
..dist_tensor
import
DistributedTensor
from
..process_group
import
get_process_group
from
..process_group
import
get_process_group
from
..utils
import
_get_comm_group
,
_get_idx_in_axis
from
..utils
import
_get_comm_group
,
_get_idx_in_axis
...
@@ -785,9 +785,12 @@ class CommOpCost(OpCost):
...
@@ -785,9 +785,12 @@ class CommOpCost(OpCost):
if
self
.
op
is
not
None
:
if
self
.
op
is
not
None
:
vars
=
self
.
op
.
block
.
vars
vars
=
self
.
op
.
block
.
vars
# NOTE: The tensor communicated input_name is "X" in default. Otherwise, this function should be overrided
# NOTE: The tensor communicated input_name is "X" in default. Otherwise, this function should be overrided
try
:
var_name
=
self
.
op
.
input
(
"X"
)[
0
]
var_name
=
self
.
op
.
input
(
"X"
)[
0
]
except
:
var_name
=
self
.
op
.
output
(
"Out"
)[
0
]
var
=
get_var_with_recursion
(
var
=
get_var_with_recursion
(
var_name
,
self
.
op
.
block
,
self
.
program
var_name
,
self
.
op
.
block
,
self
.
op
.
block
.
program
)
)
dtype
=
var
.
dtype
dtype
=
var
.
dtype
shape
=
var
.
shape
shape
=
var
.
shape
...
@@ -838,7 +841,7 @@ class CommOpCost(OpCost):
...
@@ -838,7 +841,7 @@ class CommOpCost(OpCost):
if
self
.
op_desc
is
not
None
:
if
self
.
op_desc
is
not
None
:
self
.
_group_ranks
=
self
.
op_desc
[
"group_ranks"
]
self
.
_group_ranks
=
self
.
op_desc
[
"group_ranks"
]
elif
self
.
op
is
not
None
:
elif
self
.
op
is
not
None
:
ring_id
=
self
.
op
.
attr
s
(
"ring_id"
)
ring_id
=
self
.
op
.
attr
(
"ring_id"
)
process_group
=
get_process_group
(
ring_id
)
process_group
=
get_process_group
(
ring_id
)
if
process_group
is
None
:
if
process_group
is
None
:
raise
ValueError
(
raise
ValueError
(
...
@@ -921,3 +924,57 @@ def calc_time_by_modeling(op=None, desc=None, cluster=None):
...
@@ -921,3 +924,57 @@ def calc_time_by_modeling(op=None, desc=None, cluster=None):
)
)
time
=
op_cost
.
calc_time
()
time
=
op_cost
.
calc_time
()
return
time
return
time
def
calc_time_by_cost_model
(
op
,
cluster
=
None
):
"""Calc op time by cost model and the unit is microsecond."""
if
not
isinstance
(
op
,
paddle
.
fluid
.
framework
.
Operator
):
raise
TypeError
(
"OP must be paddle.fluid.framework.Operator, but got {}."
.
format
(
type
(
op
)
)
)
if
not
cluster
:
cluster
=
get_default_cluster
()
time
=
0.0
op_type
=
op
.
type
# calc comp op time by flops
if
op_type
not
in
NON_COMP_TYPE
:
attrs
=
op
.
all_attrs
()
# build comp op inputs desc to calc flops.
# for example, a matmul op inputs desc will be {"X": [(1024, 1024)], "Y": [(1024, 1024)]}
inputs
=
{}
for
input_name
in
op
.
input_names
:
var_names
=
op
.
input
(
input_name
)
inputs
[
input_name
]
=
[]
for
var_name
in
var_names
:
var
=
op
.
block
.
_var_recursive
(
var_name
)
inputs
[
input_name
].
append
(
var
.
shape
)
# the time of grad operator is twice than its forward operator empirically
if
"_grad"
in
op_type
:
op_type
=
op_type
[:
len
(
op_type
)
-
5
]
flops_count
=
2
*
flops
(
op_type
,
inputs
,
attrs
)
else
:
flops_count
=
flops
(
op_type
,
inputs
,
attrs
)
if
cluster
.
_gpu_model
==
"V100"
:
time
=
flops_count
*
2.9e-7
*
2.6
elif
cluster
.
_gpu_model
==
"A100"
:
time
=
flops_count
*
2.9e-7
else
:
raise
ValueError
(
"Only A100 and V100 gpu has been supported currently."
)
# calc comm op time by communication modeling formula
elif
op_type
in
COMM_OP_TYPE
:
op_cost
=
_g_op_cost_factory
[
op_type
](
op
=
op
,
comm_context
=
CommContext
(
cluster
)
)
time
=
op_cost
.
calc_time
()
else
:
raise
ValueError
(
f
"The
{
op_type
}
has not been supported now."
)
return
time
test/auto_parallel/CMakeLists.txt
浏览文件 @
51ba2a0f
...
@@ -162,6 +162,7 @@ if(WITH_DISTRIBUTE AND WITH_GPU)
...
@@ -162,6 +162,7 @@ if(WITH_DISTRIBUTE AND WITH_GPU)
py_test_modules
(
test_rule_based_tuner MODULES test_rule_based_tuner
)
py_test_modules
(
test_rule_based_tuner MODULES test_rule_based_tuner
)
py_test_modules
(
test_dist_tensor MODULES test_dist_tensor
)
py_test_modules
(
test_dist_tensor MODULES test_dist_tensor
)
py_test_modules
(
test_shard_tensor_api MODULES test_shard_tensor_api
)
py_test_modules
(
test_shard_tensor_api MODULES test_shard_tensor_api
)
py_test_modules
(
test_cost_interface MODULES test_cost_interface
)
# End of unittests WITH single card WITHOUT timeout
# End of unittests WITH single card WITHOUT timeout
endif
()
endif
()
test/auto_parallel/test_cost_interface.py
0 → 100644
浏览文件 @
51ba2a0f
# Copyright (c) 2021 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
unittest
import
paddle
import
paddle.nn.functional
as
F
from
paddle
import
nn
,
static
,
utils
from
paddle.distributed
import
fleet
from
paddle.distributed.auto_parallel.static.cluster
import
Cluster
from
paddle.distributed.auto_parallel.static.completion
import
Completer
from
paddle.distributed.auto_parallel.static.cost
import
calc_time_by_cost_model
from
paddle.distributed.auto_parallel.static.dist_context
import
(
DistributedContext
,
)
from
paddle.distributed.auto_parallel.static.parallelizer
import
(
AutoParallelizer
,
)
from
paddle.distributed.auto_parallel.static.partitioner
import
Partitioner
from
paddle.distributed.auto_parallel.static.reshard
import
Resharder
from
paddle.distributed.fleet
import
auto
paddle
.
enable_static
()
_global_parallel_strategy
=
"dp_mp_pp"
_global_process_mesh
=
auto
.
ProcessMesh
(
[[[
0
,
1
],
[
4
,
5
]],
[[
2
,
3
],
[
6
,
7
]]],
dim_names
=
[
"x"
,
"y"
,
"z"
]
)
PP_MESH_0
=
auto
.
ProcessMesh
([[
0
,
1
],
[
4
,
5
]],
dim_names
=
[
"x"
,
"y"
])
PP_MESH_1
=
auto
.
ProcessMesh
([[
2
,
3
],
[
6
,
7
]],
dim_names
=
[
"x"
,
"y"
])
class
MLPLayer
(
nn
.
Layer
):
def
__init__
(
self
,
hidden_size
=
1024
,
intermediate_size
=
4
*
1024
,
initializer_range
=
0.02
,
):
super
().
__init__
()
d_model
=
hidden_size
dim_feedforward
=
intermediate_size
weight_attr
=
paddle
.
ParamAttr
(
initializer
=
nn
.
initializer
.
Normal
(
mean
=
0.0
,
std
=
initializer_range
)
)
bias_attr
=
None
self
.
linear0
=
nn
.
Linear
(
d_model
,
dim_feedforward
,
weight_attr
,
bias_attr
=
bias_attr
)
self
.
linear1
=
nn
.
Linear
(
dim_feedforward
,
d_model
,
weight_attr
,
bias_attr
=
bias_attr
)
self
.
norm
=
nn
.
LayerNorm
(
d_model
,
epsilon
=
1e-5
)
def
forward
(
self
,
input
):
auto
.
shard_tensor
(
self
.
linear0
.
weight
,
PP_MESH_0
,
[
None
,
"y"
])
auto
.
shard_tensor
(
self
.
linear1
.
weight
,
PP_MESH_1
,
[
"y"
,
None
])
out
=
self
.
norm
(
input
)
out
=
self
.
linear0
(
out
)
out
=
F
.
gelu
(
out
,
approximate
=
True
)
out
=
self
.
linear1
(
out
)
param
=
paddle
.
create_parameter
([
1024
,
4096
],
paddle
.
float32
)
auto
.
shard_tensor
(
param
,
PP_MESH_1
,
[
None
,
"y"
])
out
=
paddle
.
matmul
(
out
,
param
)
return
out
def
mlp_forward
(
train_program
,
start_program
):
with
static
.
program_guard
(
train_program
,
start_program
),
utils
.
unique_name
.
guard
():
batch_size
=
4
hidden_size
=
1024
sequence_len
=
512
input
=
static
.
data
(
name
=
"input"
,
shape
=
[
batch_size
,
hidden_size
],
dtype
=
'float32'
)
label
=
static
.
data
(
name
=
"label"
,
shape
=
[
batch_size
,
1
],
dtype
=
'float32'
)
auto
.
shard_tensor
(
input
,
PP_MESH_0
,
[
"x"
,
None
])
auto
.
shard_tensor
(
label
,
PP_MESH_1
,
[
"x"
,
None
])
mlp
=
MLPLayer
(
hidden_size
=
hidden_size
,
intermediate_size
=
4
*
hidden_size
,
initializer_range
=
0.02
,
)
predict
=
mlp
(
input
)
error_cost
=
paddle
.
nn
.
functional
.
square_error_cost
(
predict
,
label
)
loss
=
paddle
.
mean
(
error_cost
)
return
loss
,
train_program
,
start_program
def
get_dist_prog
(
train_program
,
startup_program
,
dist_context
,
rank_id
):
global
_global_process_mesh
dist_context
.
process_mesh
=
_global_process_mesh
loss
,
train_program
,
startup_program
=
mlp_forward
(
train_program
,
startup_program
)
fleet
.
_user_defined_strategy
=
fleet
.
DistributedStrategy
()
fleet
.
user_defined_optimizer
=
paddle
.
optimizer
.
Adam
()
parallelizer
=
AutoParallelizer
(
fleet
)
parallelizer
.
_dist_context
=
dist_context
# serial forward & backward completion
completer
=
Completer
(
dist_context
)
complete_train_program
=
completer
.
complete_forward_annotation
(
train_program
)
dist_context
.
block_state
.
parse_forward_blocks
(
complete_train_program
)
params_grads
=
parallelizer
.
_generate_backward
(
complete_train_program
,
startup_program
,
loss
,
parameter_list
=
None
,
no_grad_set
=
None
,
callbacks
=
None
,
)
# logical partition
partitioner
=
Partitioner
(
dist_context
,
rank_id
)
(
auto_parallel_main_prog
,
auto_parallel_startup_prog
,
dist_params_grads
,
)
=
partitioner
.
partition
(
complete_train_program
,
startup_program
,
params_grads
)
partitioned_optimize_ops
=
parallelizer
.
_apply_optimize
(
auto_parallel_main_prog
,
auto_parallel_startup_prog
,
dist_params_grads
)
return
(
auto_parallel_main_prog
,
auto_parallel_startup_prog
,
dist_params_grads
,
)
class
TestCostInterface
(
unittest
.
TestCase
):
def
test_cost_interface
(
self
):
train_program
=
paddle
.
static
.
Program
()
startup_program
=
paddle
.
static
.
Program
()
dist_context
=
DistributedContext
()
rank_id
=
2
dist_main_prog
,
dist_startup_prog
,
dist_params_grads
=
get_dist_prog
(
train_program
,
startup_program
,
dist_context
,
rank_id
)
resharder
=
Resharder
(
dist_main_prog
,
dist_startup_prog
,
rank_id
,
dist_context
,
dist_params_grads
,
)
resharder
.
reshard
()
cluster
=
Cluster
()
cluster
.
gen_default_config_cluster
(
node_count
=
1
,
device_count
=
8
)
for
op
in
dist_main_prog
.
global_block
().
ops
:
time
=
calc_time_by_cost_model
(
op
,
cluster
)
assert
time
>
-
1
if
__name__
==
"__main__"
:
unittest
.
main
()
编辑
预览
Markdown
is supported
0%
请重试
或
添加新附件
.
添加附件
取消
You are about to add
0
people
to the discussion. Proceed with caution.
先完成此消息的编辑!
取消
想要评论请
注册
或
登录