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
BaiXuePrincess
Paddle
提交
b54d1ba9
P
Paddle
项目概览
BaiXuePrincess
/
Paddle
与 Fork 源项目一致
Fork自
PaddlePaddle / Paddle
通知
1
Star
1
Fork
0
代码
文件
提交
分支
Tags
贡献者
分支图
Diff
Issue
0
列表
看板
标记
里程碑
合并请求
0
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
0
Issue
0
列表
看板
标记
里程碑
合并请求
0
合并请求
0
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
提交
b54d1ba9
编写于
6年前
作者:
Y
Yancey1989
浏览文件
操作
浏览文件
下载
电子邮件补丁
差异文件
fix pserver sub-blocks
上级
59729902
develop
2.0.1-rocm-post
Ligoml-patch-1
OliverLPH-patch-1
OliverLPH-patch-2
PaddlePM-patch-1
PaddlePM-patch-2
ZHUI-patch-1
add_default_att
add_model_benchmark_ci
add_some_yaml_config
addfile
all_new_design_exec
ascendrc
ascendrelease
cherry_undefined_var
compile_windows
cp_2.4_fix_numpy
delete_2.0.1-rocm-post
delete_add_default_att
delete_all_new_design_exec
delete_ascendrc
delete_compile_windows
delete_delete_addfile
delete_disable_iterable_dataset_unittest
delete_fix_dataloader_memory_leak
delete_fix_imperative_dygraph_error
delete_fix_retry_ci
delete_fix_undefined_var
delete_improve_sccache
delete_incubate/lite
delete_paddle_tiny_install
delete_paralleltest
delete_prv-disable-more-cache
delete_revert-31068-fix_conv3d_windows
delete_revert-31562-mean
delete_revert-33630-bug-fix
delete_revert-34159-add_npu_bce_logical_dev
delete_revert-34910-spinlocks_for_allocator
delete_revert-35069-revert-34910-spinlocks_for_allocator
delete_revert-36057-dev/read_flags_in_ut
dingjiaweiww-patch-1
disable_iterable_dataset_unittest
dy2static
enable_eager_model_test
final_state_gen_python_c
final_state_intermediate
fix-numpy-issue
fix_concat_slice
fix_dataloader_memory_leak
fix_dlpack_for
fix_imperative_dygraph_error
fix_npu_ci
fix_op_flops
fix_retry_ci
fix_rnn_docs
fix_tensor_type
fix_undefined_var
fix_var_stop_gradient_error
fixiscan
fixiscan1
fixiscan2
fixiscan3
github/fork/123malin/netifaces
github/fork/123malin/tdm_abacus
github/fork/AshburnLee/dev_unique
github/fork/ForFishes/fix_memory_matmul
github/fork/ForFishes/rm_fluid
github/fork/LielinJiang/move-2.0-api
github/fork/LielinJiang/visual-dl-cb
github/fork/LiuChiachi/add-transformer-generate-square-subsequent-mask-api
github/fork/LiuChiachi/fix-example-code-for-hapi-Model
github/fork/LiuChiachi/remove-input-requirment-in-dygraph-Model
github/fork/MrChengmo/fix_ps_profiler
github/fork/MrChengmo/update_ps_heter
github/fork/PWhiddy/patch-1
github/fork/Shixiaowei02/dev/save_load_upgrade
github/fork/TCChenlong/fix_hapi
github/fork/TCChenlong/fix_inden
github/fork/Thunderbrook/xpu_slice
github/fork/XieYunshen/disable_ut_test_parallel_executor_fetch_isolated_var
github/fork/XieYunshen/disable_ut_test_parallel_executor_fetch_isolated_var_2
github/fork/XieYunshen/disable_ut_test_parallel_executor_fetch_isolated_var_3
github/fork/XieYunshen/timeout_20S_ut
github/fork/ZeyuChen/remove-nltk
github/fork/arlesniak/arlesniak/selective__mkldnn_flags
github/fork/baiyfbupt/code_doc_mig
github/fork/chalsliu/set_timeout
github/fork/chen-zhiyu/develop
github/fork/chenwhql/ci/try_to_find_test_buffer_shared_memory_reuse_pass_error
github/fork/chenwhql/dygraph/remove_scale_loss_and_apply_collective_grads
github/fork/chenwhql/saveload/add_get_inference_program
github/fork/chenwhql/saveload/remove_save_load_config
github/fork/cryoco/pass-compatibility-trt
github/fork/danleifeng/isempty_api2.0
github/fork/frankwhzhang/api_transfer
github/fork/hbwx24/error_msg/cuda_kernel_error_msg
github/fork/heavengate/cherry_yolo_box
github/fork/heavengate/update_yolo_box
github/fork/iclementine/rnn_fix
github/fork/iducn/testestse
github/fork/jczaja/prv-25537-fix
github/fork/jeff41404/release/1.8
github/fork/jiweibo/api_2.0
github/fork/jiweibo/fix_lite_resnet50_test
github/fork/juncaipeng/fix_doc_1
github/fork/lfchener/sample_code
github/fork/littletomatodonkey/fix_reg_doc
github/fork/liym27/dy2stat_update_assign_to_rc20
github/fork/luotao1/profiler_ut
github/fork/mapingshuo/add_wait
github/fork/mapingshuo/doc_2.0
github/fork/mapingshuo/zero-0.5
github/fork/miraiwk/dev
github/fork/pangyoki/add-Categorical-class-branch
github/fork/pangyoki/add-multinomial-op-branch
github/fork/pangyoki/fix-test_distritbution-CI
github/fork/qjing666/doublegrad
github/fork/qjing666/fix_hdfs_download
github/fork/sandyhouse/add_gather_etc
github/fork/sandyhouse/add_send_recv_alltoall_etc
github/fork/sandyhouse/pipeline_exe_run
github/fork/seiriosPlus/feature/large_scale_kv_save_delta
github/fork/seiriosPlus/fix/paddle_errors_fix
github/fork/seiriosPlus/fix/paddle_op_errors
github/fork/shangzhizhou/fix_test_activation_op_random_bug
github/fork/smallv0221/yxp0924
github/fork/smallv0221/yxp0925
github/fork/swtkiwi/del-matplotlib
github/fork/tianshuo78520a/kunlun_test
github/fork/tianshuo78520a/update_dockerfile
github/fork/wanghaoshuang/bert_fuse
github/fork/wanghaoshuang/label_smooth
github/fork/wanghuancoder/develop_CUDASynchronize
github/fork/wanghuancoder/develop_Layer_doc
github/fork/wanghuancoder/develop_ParameterList_doc
github/fork/wanghuancoder/develop_Sequential_doc
github/fork/wanghuancoder/develop_bilinear_tensor_product
github/fork/wanghuancoder/develop_coverage_build_sh
github/fork/wanghuancoder/develop_in_dynamic_mode_doc
github/fork/wanghuancoder/develop_unique_name_doc
github/fork/wangxicoding/fleet_meta_combine
github/fork/wawltor/error_message_fix_5
github/fork/willthefrog/remove_l2_norm
github/fork/windstamp/momentum_op
github/fork/windstamp/mv_op_5
github/fork/windstamp/normal_api
github/fork/wojtuss/wojtuss/fusion_gru_quantization
github/fork/wojtuss/wojtuss/quantization-with-shift
github/fork/wzzju/fix_err_info
github/fork/wzzju/pure_fp16
github/fork/xiemoyuan/op_error_message
github/fork/xiemoyuan/optimize_error_message
github/fork/yaoxuefeng6/fix_doc
github/fork/yaoxuefeng6/mod_dataset_v2
github/fork/yongqiangma/lod
github/fork/ysh329/fix-clip-by-norm-error
github/fork/ysh329/fix-error-clip-by-value
github/fork/yukavio/error_info
github/fork/zhangting2020/conv_filter_grad
github/fork/zhangting2020/is_compile_with_cuda
github/fork/zhangting2020/place_doc
github/fork/zhangting2020/program
github/fork/zhhsplendid/fix_any
github/fork/zhhsplendid/refine_api2
github/fork/zhhsplendid/refine_api2_test
github/fork/zhhsplendid/refine_api_test_ptb_lm
github/fork/zhhsplendid/refine_api_test_resnet
github/fork/zhhsplendid/refine_api_test_simnet
github/fork/zhiqiu/dev/refine_initializer
github/fork/zhiqiu/dev/remove_inplace_argument
github/fork/zlsh80826/nvinfer_plugin_var_len_cuda11
improve_sccache
incubate/frl_train_eval
incubate/infrt
incubate/lite
inplace_addto
layer_norm
make_flag_adding_easier
matmul_double_grad
move_embedding_to_phi
move_histogram_to_pten
move_sgd_to_phi
move_slice_to_pten
move_temporal_shift_to_phi
move_yolo_box_to_phi
npu_fix_alloc
numel
paddle_tiny_install
paralleltest
preln_ernie
prv-disable-more-cache
prv-md-even-more
prv-onednn-2.5
prv-reshape-mkldnn-ut2
pten_tensor_refactor
release/0.15.0
release/1.0.0
release/1.1
release/1.2
release/1.3
release/1.4
release/1.5
release/1.6
release/1.7
release/1.8
release/2.0
release/2.0-alpha
release/2.0-beta
release/2.0-rc
release/2.0-rc1
release/2.1
release/2.2
release/2.3
release/2.3-fc-ernie-fix
release/2.4
release/lite-0.1
revert-24981-add_device_attr_for_regulization
revert-26856-strategy_example2
revert-27520-disable_pr
revert-31068-fix_conv3d_windows
revert-31562-mean
revert-32290-develop-hardlabel
revert-33037-forci
revert-33475-fix_cifar_label_dimension
revert-33630-bug-fix
revert-34159-add_npu_bce_logical_dev
revert-34406-add_copy_from_tensor
revert-34910-spinlocks_for_allocator
revert-35069-revert-34910-spinlocks_for_allocator
revert-36057-dev/read_flags_in_ut
revert-36201-refine_fast_threaded_ssa_graph_executor
revert-36985-add_license
revert-37318-refactor_dygraph_to_eager
revert-37926-eager_coreops_500
revert-37956-revert-37727-pylayer_support_tuple
revert-38100-mingdong
revert-38301-allocation_rearrange_pr
revert-38703-numpy_bf16_package_reupload
revert-38732-remove_useless_header_in_elementwise_mul_grad
revert-38959-Reduce_Grad
revert-39143-adjust_empty
revert-39227-move_trace_op_to_pten
revert-39268-dev/remove_concat_fluid_kernel
revert-40170-support_partial_grad
revert-41056-revert-40727-move_some_activaion_to_phi
revert-41065-revert-40993-mv_ele_floordiv_pow
revert-41068-revert-40790-phi_new
revert-41944-smaller_inference_api_test
revert-42149-do-not-reset-default-stream-for-stream-safe-cuda-allocator
revert-43155-fix_ut_tempfile
revert-43882-revert-41944-smaller_inference_api_test
revert-45808-phi/simplify_size_op
revert-46827-deform_comment
revert-47325-remove_cudnn_hardcode
revert-47645-add_npu_storage_dims
revert-48815-set_free_when_no_cache_hit_default_value_true
revert-49654-prim_api_gen
revert-49763-fix_static_composite_gen
rocm_dev_0217
support-0D-sort
support_weight_transpose
test_benchmark_ci
test_feature_precision_test_c
test_for_Filtetfiles
test_model_benchmark
test_model_benchmark_ci
zhiqiu-patch-1
v2.4.1
v2.4.0
v2.4.0-rc0
v2.3.2
v2.3.1
v2.3.0
v2.3.0-rc0
v2.2.2
v2.2.1
v2.2.0
v2.2.0-rc0
v2.2.0-bak0
v2.1.3
v2.1.2
v2.1.1
v2.1.0
v2.1.0-rc0
v2.0.2
v2.0.1
v2.0.0
v2.0.0-rc1
v2.0.0-rc0
v2.0.0-beta0
v2.0.0-alpha0
v1.8.5
v1.8.4
v1.8.3
v1.8.2
v1.8.1
v1.8.0
v1.7.2
v1.7.1
v1.7.0
v1.6.3
v1.6.2
v1.6.1
v1.6.0
v1.6.0-rc0
v1.5.2
v1.5.1
v1.5.0
v1.4.1
v1.4.0
v1.3.2
v1.3.1
v1.3.0
v1.2.1
v1.2.0
v1.1.0
v1.0.2
v1.0.1
v1.0.0
v1.0.0-rc0
v0.15.0
v0.15.0-rc0
lite-v0.1
无相关合并请求
变更
2
显示空白变更内容
内联
并排
Showing
2 changed file
with
39 addition
and
20 deletion
+39
-20
paddle/fluid/operators/listen_and_serv_op.cc
paddle/fluid/operators/listen_and_serv_op.cc
+9
-1
python/paddle/fluid/transpiler/distribute_transpiler.py
python/paddle/fluid/transpiler/distribute_transpiler.py
+30
-19
未找到文件。
paddle/fluid/operators/listen_and_serv_op.cc
浏览文件 @
b54d1ba9
...
@@ -101,13 +101,16 @@ void ListenAndServOp::RunSyncLoop(
...
@@ -101,13 +101,16 @@ void ListenAndServOp::RunSyncLoop(
framework
::
Scope
*
recv_scope
,
framework
::
Scope
*
recv_scope
,
const
std
::
vector
<
int
>
&
prefetch_block_id_list
)
const
{
const
std
::
vector
<
int
>
&
prefetch_block_id_list
)
const
{
size_t
num_blocks
=
program
->
Size
();
size_t
num_blocks
=
program
->
Size
();
auto
skip_sub_blks
=
Attr
<
std
::
vector
<
int
>>
(
"skip_sub_blks"
);
PADDLE_ENFORCE_GE
(
num_blocks
,
2
,
PADDLE_ENFORCE_GE
(
num_blocks
,
2
,
"server program should have at least 2 blocks"
);
"server program should have at least 2 blocks"
);
std
::
vector
<
int
>
optimize_block_id_list
;
std
::
vector
<
int
>
optimize_block_id_list
;
for
(
int
blkid
=
1
;
blkid
<
num_blocks
;
++
blkid
)
{
for
(
int
blkid
=
1
;
blkid
<
num_blocks
;
++
blkid
)
{
if
(
std
::
find
(
prefetch_block_id_list
.
begin
(),
prefetch_block_id_list
.
end
(),
if
(
std
::
find
(
prefetch_block_id_list
.
begin
(),
prefetch_block_id_list
.
end
(),
blkid
)
==
prefetch_block_id_list
.
end
())
{
blkid
)
==
prefetch_block_id_list
.
end
()
&&
std
::
find
(
skip_sub_blks
.
begin
(),
skip_sub_blks
.
end
(),
blkid
)
==
skip_sub_blks
.
end
())
{
optimize_block_id_list
.
push_back
(
blkid
);
optimize_block_id_list
.
push_back
(
blkid
);
}
}
}
}
...
@@ -344,6 +347,11 @@ class ListenAndServOpMaker : public framework::OpProtoAndCheckerMaker {
...
@@ -344,6 +347,11 @@ class ListenAndServOpMaker : public framework::OpProtoAndCheckerMaker {
.
SetDefault
({});
.
SetDefault
({});
AddAttr
<
int
>
(
"Fanin"
,
"How many clients send to this server."
)
AddAttr
<
int
>
(
"Fanin"
,
"How many clients send to this server."
)
.
SetDefault
(
1
);
.
SetDefault
(
1
);
AddAttr
<
std
::
vector
<
int
>>
(
"skip_sub_blks"
,
"do not parallel execute the specify sub blocks, "
"it's used for the op which has"
"condition blocks"
)
.
SetDefault
({});
}
}
};
};
...
...
This diff is collapsed.
Click to expand it.
python/paddle/fluid/transpiler/distribute_transpiler.py
浏览文件 @
b54d1ba9
...
@@ -250,19 +250,14 @@ class DistributeTranspiler:
...
@@ -250,19 +250,14 @@ class DistributeTranspiler:
split_method
=
RoundRobin
,
split_method
=
RoundRobin
,
sync_mode
=
True
):
sync_mode
=
True
):
"""
"""
:param trainer_id: one unique id for each trainer in a job.
Args:
:type trainer_id: int
trainer_id(int): one unique id for each trainer in a job.
:param program: program to transpile, default is default_main_program
program(Program): program to transpile, default is default_main_program
:type program: Program
pservers(string): parameter server endpoints like "m1:6174,m2:6174"
:param pservers: parameter server endpoints like "m1:6174,m2:6174"
trainers(int): total number of workers/trainers in the job
:type pservers: string
split_method(PSDispatcher): A function to determin how to split variables
:param trainers: total number of workers/trainers in the job
:type trainers: int
:param split_method: A function to determin how to split variables
to different servers equally.
to different servers equally.
:type split_method: function
sync_mode(boolean): if sync_mode is set True, it means that dist transpiler
:type sync_mode: boolean default True
:param sync_mode: if sync_mode is set True, it means that dist transpiler
will transpile the program into sync_mode pserver and trainer program.
will transpile the program into sync_mode pserver and trainer program.
"""
"""
assert
(
split_method
.
__bases__
[
0
]
==
PSDispatcher
)
assert
(
split_method
.
__bases__
[
0
]
==
PSDispatcher
)
...
@@ -403,6 +398,11 @@ class DistributeTranspiler:
...
@@ -403,6 +398,11 @@ class DistributeTranspiler:
NOTE: assume blocks of the same variable is not distributed
NOTE: assume blocks of the same variable is not distributed
on the same pserver, only change param/grad varnames for
on the same pserver, only change param/grad varnames for
trainers to fetch.
trainers to fetch.
Args:
endpoint(string): the endpoint for the current pserver instance.
Returns(Program): the pserver program
"""
"""
# step1
# step1
pserver_program
=
Program
()
pserver_program
=
Program
()
...
@@ -479,9 +479,9 @@ class DistributeTranspiler:
...
@@ -479,9 +479,9 @@ class DistributeTranspiler:
return
varname
return
varname
return
""
return
""
def
__clone_lr_op_sub_block__
(
op
,
program
,
new_block
):
def
__clone_lr_op_sub_block__
(
op
,
program
,
new_block
,
skip_sub_blks
):
if
not
op
.
has_attr
(
'sub_block'
):
if
not
op
.
has_attr
(
'sub_block'
):
return
return
-
1
origin_block_desc
=
op
.
attr
(
'sub_block'
)
origin_block_desc
=
op
.
attr
(
'sub_block'
)
origin_block
=
self
.
origin_program
.
block
(
origin_block_desc
.
id
)
origin_block
=
self
.
origin_program
.
block
(
origin_block_desc
.
id
)
...
@@ -489,6 +489,7 @@ class DistributeTranspiler:
...
@@ -489,6 +489,7 @@ class DistributeTranspiler:
# we put the new sub block to new block to follow the block
# we put the new sub block to new block to follow the block
# hierarchy of the original blocks
# hierarchy of the original blocks
new_sub_block
=
program
.
create_block
(
new_block
.
idx
)
new_sub_block
=
program
.
create_block
(
new_block
.
idx
)
skip_sub_blks
(
new_sub_block
.
idx
)
# clone vars
# clone vars
for
var
in
origin_block
.
vars
:
for
var
in
origin_block
.
vars
:
...
@@ -498,20 +499,24 @@ class DistributeTranspiler:
...
@@ -498,20 +499,24 @@ class DistributeTranspiler:
for
op
in
origin_block
.
ops
:
for
op
in
origin_block
.
ops
:
self
.
_clone_lr_op
(
program
,
new_sub_block
,
op
)
self
.
_clone_lr_op
(
program
,
new_sub_block
,
op
)
# clone sub_block of op
# clone sub_block of op
__clone_lr_op_sub_block__
(
op
,
program
,
new_sub_block
)
__clone_lr_op_sub_block__
(
op
,
program
,
new_sub_block
,
skip_sub_blks
)
# reset the block of op
# reset the block of op
op
.
set_attr
(
'sub_block'
,
new_sub_block
)
op
.
set_attr
(
'sub_block'
,
new_sub_block
)
return
new_sub_block
.
idx
# append lr decay ops to the child block if exists
# append lr decay ops to the child block if exists
lr_ops
=
self
.
_get_lr_ops
()
lr_ops
=
self
.
_get_lr_ops
()
skip_sub_blks
=
[]
if
len
(
lr_ops
)
>
0
:
if
len
(
lr_ops
)
>
0
:
lr_decay_block
=
pserver_program
.
create_block
(
lr_decay_block
=
pserver_program
.
create_block
(
pserver_program
.
num_blocks
-
1
)
pserver_program
.
num_blocks
-
1
)
for
_
,
op
in
enumerate
(
lr_ops
):
for
_
,
op
in
enumerate
(
lr_ops
):
self
.
_append_pserver_non_opt_ops
(
lr_decay_block
,
op
)
self
.
_append_pserver_non_opt_ops
(
lr_decay_block
,
op
)
# append sub blocks to pserver_program in lr_decay_op
# append sub blocks to pserver_program in lr_decay_op
__clone_lr_op_sub_block__
(
op
,
pserver_program
,
lr_decay_block
)
__clone_lr_op_sub_block__
(
op
,
pserver_program
,
lr_decay_block
,
skip_sub_blks
)
# append op to the current block
# append op to the current block
grad_to_block_id
=
[]
grad_to_block_id
=
[]
...
@@ -561,7 +566,8 @@ class DistributeTranspiler:
...
@@ -561,7 +566,8 @@ class DistributeTranspiler:
"endpoint"
:
endpoint
,
"endpoint"
:
endpoint
,
"Fanin"
:
self
.
trainer_num
,
"Fanin"
:
self
.
trainer_num
,
"sync_mode"
:
self
.
sync_mode
,
"sync_mode"
:
self
.
sync_mode
,
"grad_to_block_id"
:
grad_to_block_id
"grad_to_block_id"
:
grad_to_block_id
,
"skip_sub_blks"
:
skip_sub_blks
}
}
if
len
(
prefetch_var_name_to_block_id
)
>
0
:
if
len
(
prefetch_var_name_to_block_id
)
>
0
:
attrs
[
'prefetch_var_name_to_block_id'
]
\
attrs
[
'prefetch_var_name_to_block_id'
]
\
...
@@ -582,6 +588,11 @@ class DistributeTranspiler:
...
@@ -582,6 +588,11 @@ class DistributeTranspiler:
Get startup program for current parameter server.
Get startup program for current parameter server.
Modify operator input variables if there are variables that
Modify operator input variables if there are variables that
were split to several blocks.
were split to several blocks.
Args:
endpoint(string): the endpoint for the current pserver instance.
pserver_program(Program): the program for pserver to execute.
Returns(Program): the startup program for pserver
"""
"""
s_prog
=
Program
()
s_prog
=
Program
()
orig_s_prog
=
default_startup_program
()
orig_s_prog
=
default_startup_program
()
...
...
This diff is collapsed.
Click to expand it.
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
反馈
建议
客服
返回
顶部