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
Crayon鑫
Paddle
提交
1edc0423
P
Paddle
项目概览
Crayon鑫
/
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
1
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
1
Issue
1
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
1edc0423
编写于
1月 24, 2019
作者:
Q
Qiao Longfei
浏览文件
操作
浏览文件
下载
电子邮件补丁
差异文件
update send_op
上级
74040cb4
变更
1
隐藏空白更改
内联
并排
Showing
1 changed file
with
38 addition
and
21 deletion
+38
-21
paddle/fluid/operators/distributed_ops/send_op.cc
paddle/fluid/operators/distributed_ops/send_op.cc
+38
-21
未找到文件。
paddle/fluid/operators/distributed_ops/send_op.cc
浏览文件 @
1edc0423
...
@@ -20,6 +20,7 @@ limitations under the License. */
...
@@ -20,6 +20,7 @@ limitations under the License. */
#include "paddle/fluid/framework/lod_tensor.h"
#include "paddle/fluid/framework/lod_tensor.h"
#include "paddle/fluid/framework/op_registry.h"
#include "paddle/fluid/framework/op_registry.h"
#include "paddle/fluid/operators/distributed/distributed.h"
#include "paddle/fluid/operators/distributed/distributed.h"
#include "paddle/fluid/operators/distributed/parameter_send.h"
#include "paddle/fluid/operators/distributed_ops/send_recv_util.h"
#include "paddle/fluid/operators/distributed_ops/send_recv_util.h"
#include "paddle/fluid/platform/profiler.h"
#include "paddle/fluid/platform/profiler.h"
...
@@ -37,30 +38,46 @@ class SendOp : public framework::OperatorBase {
...
@@ -37,30 +38,46 @@ class SendOp : public framework::OperatorBase {
const
platform
::
Place
&
place
)
const
override
{
const
platform
::
Place
&
place
)
const
override
{
auto
ins
=
Inputs
(
"X"
);
auto
ins
=
Inputs
(
"X"
);
std
::
vector
<
std
::
string
>
epmap
=
Attr
<
std
::
vector
<
std
::
string
>>
(
"epmap"
);
auto
epmap
=
Attr
<
std
::
vector
<
std
::
string
>>
(
"epmap"
);
int
sync_send
=
Attr
<
int
>
(
"sync_mode"
);
int
sync_send
=
Attr
<
int
>
(
"sync_mode"
);
platform
::
DeviceContextPool
&
pool
=
platform
::
DeviceContextPool
::
Instance
();
auto
send_varnames
=
Attr
<
std
::
vector
<
std
::
string
>>
(
"send_varnames"
);
auto
&
ctx
=
*
pool
.
Get
(
place
);
auto
height_sections
=
Attr
<
std
::
vector
<
int64_t
>>
(
"height_sections"
);
distributed
::
RPCClient
*
rpc_client
=
if
(
send_varnames
.
size
()
>
0
)
{
distributed
::
RPCClient
::
GetInstance
<
RPCCLIENT_T
>
(
PADDLE_ENFORCE_EQ
(
ins
.
size
(),
1
,
""
);
Attr
<
int
>
(
"trainer_id"
));
framework
::
RuntimeContext
ctx
(
Inputs
(),
Outputs
(),
scope
);
platform
::
DeviceContextPool
&
pool
=
std
::
vector
<
distributed
::
VarHandlePtr
>
rets
;
platform
::
DeviceContextPool
::
Instance
();
for
(
size_t
i
=
0
;
i
<
ins
.
size
();
i
++
)
{
auto
*
dev_ctx
=
pool
.
Get
(
place
);
if
(
NeedSend
(
scope
,
ins
[
i
]))
{
auto
exe_ctx
=
framework
::
ExecutionContext
(
*
this
,
scope
,
*
dev_ctx
,
ctx
);
VLOG
(
3
)
<<
"sending "
<<
ins
[
i
]
<<
" to "
<<
epmap
[
i
];
distributed
::
send
<
float
>
(
ins
[
0
],
send_varnames
,
epmap
,
height_sections
,
rets
.
push_back
(
rpc_client
->
AsyncSendVar
(
epmap
[
i
],
ctx
,
scope
,
ins
[
i
]));
exe_ctx
,
scope
,
static_cast
<
bool
>
(
sync_send
));
}
else
{
}
else
{
VLOG
(
3
)
<<
"don't send no-initialied variable: "
<<
ins
[
i
];
platform
::
DeviceContextPool
&
pool
=
platform
::
DeviceContextPool
::
Instance
();
auto
&
ctx
=
*
pool
.
Get
(
place
);
distributed
::
RPCClient
*
rpc_client
=
distributed
::
RPCClient
::
GetInstance
<
RPCCLIENT_T
>
(
Attr
<
int
>
(
"trainer_id"
));
std
::
vector
<
distributed
::
VarHandlePtr
>
rets
;
for
(
size_t
i
=
0
;
i
<
ins
.
size
();
i
++
)
{
if
(
NeedSend
(
scope
,
ins
[
i
]))
{
VLOG
(
3
)
<<
"sending "
<<
ins
[
i
]
<<
" to "
<<
epmap
[
i
];
rets
.
push_back
(
rpc_client
->
AsyncSendVar
(
epmap
[
i
],
ctx
,
scope
,
ins
[
i
]));
}
else
{
VLOG
(
3
)
<<
"don't send no-initialied variable: "
<<
ins
[
i
];
}
}
}
}
if
(
sync_send
)
{
if
(
sync_send
)
{
for
(
size_t
i
=
0
;
i
<
rets
.
size
();
i
++
)
{
for
(
size_t
i
=
0
;
i
<
rets
.
size
();
i
++
)
{
VLOG
(
7
)
<<
"before sync_send "
<<
ins
[
i
]
<<
"from "
<<
epmap
[
i
];
VLOG
(
7
)
<<
"before sync_send "
<<
ins
[
i
]
<<
"from "
<<
epmap
[
i
]
;
PADDLE_ENFORCE
(
rets
[
i
]
->
Wait
(),
"internal error in RPCClient"
)
;
PADDLE_ENFORCE
(
rets
[
i
]
->
Wait
(),
"internal error in RPCClient"
)
;
VLOG
(
7
)
<<
"after sync_send "
<<
ins
[
i
]
<<
"from "
<<
epmap
[
i
]
;
VLOG
(
7
)
<<
"after sync_send "
<<
ins
[
i
]
<<
"from "
<<
epmap
[
i
];
}
}
}
}
}
}
}
...
...
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