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
OPTHREE
Paddle
提交
f9974a4a
P
Paddle
项目概览
OPTHREE
/
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
体验新版 GitCode,发现更多精彩内容 >>
提交
f9974a4a
编写于
3月 13, 2018
作者:
Y
Yu Yang
浏览文件
操作
浏览文件
下载
电子邮件补丁
差异文件
Make double_buffer reader async
上级
a8c076e5
变更
1
隐藏空白更改
内联
并排
Showing
1 changed file
with
42 addition
and
17 deletion
+42
-17
paddle/fluid/operators/reader/create_double_buffer_reader_op.cc
.../fluid/operators/reader/create_double_buffer_reader_op.cc
+42
-17
未找到文件。
paddle/fluid/operators/reader/create_double_buffer_reader_op.cc
浏览文件 @
f9974a4a
...
...
@@ -24,15 +24,31 @@ static constexpr size_t kDoubleBufferSize = 2;
class
DoubleBufferReader
:
public
framework
::
DecoratedReader
{
public:
struct
Item
{
Item
()
:
ctx_
(
nullptr
)
{}
std
::
vector
<
framework
::
LoDTensor
>
payloads_
;
platform
::
DeviceContext
*
ctx_
;
};
explicit
DoubleBufferReader
(
ReaderBase
*
reader
,
platform
::
Place
target_place
=
platform
::
CPUPlace
())
:
DecoratedReader
(
reader
),
place_
(
target_place
)
{
for
(
size_t
i
=
0
;
i
<
kDoubleBufferSize
;
++
i
)
{
if
(
platform
::
is_gpu_place
(
place_
))
{
#ifdef PADDLE_WITH_CUDA
ctxs_
.
emplace_back
(
new
platform
::
CUDADeviceContext
(
boost
::
get
<
platform
::
CUDAPlace
>
(
place_
)));
#else
#endif
}
}
start_thread
();
}
void
start_thread
()
{
buffer_
=
framework
::
MakeChannel
<
std
::
vector
<
framework
::
LoDTensor
>>
(
kDoubleBufferSize
);
buffer_
=
framework
::
MakeChannel
<
Item
>
(
kDoubleBufferSize
);
std
::
thread
prefetch
([
this
]
{
PrefetchThreadFunc
();
});
prefetch
.
detach
();
}
...
...
@@ -47,9 +63,10 @@ class DoubleBufferReader : public framework::DecoratedReader {
private:
void
PrefetchThreadFunc
();
framework
::
Channel
<
std
::
vector
<
framework
::
LoDTensor
>
>*
buffer_
;
framework
::
Channel
<
Item
>*
buffer_
;
platform
::
Place
place_
;
mutable
std
::
vector
<
framework
::
LoDTensor
>
local_buffer_
;
std
::
vector
<
std
::
unique_ptr
<
platform
::
DeviceContext
>>
ctxs_
;
mutable
Item
local_buffer_
;
};
class
CreateDoubleBufferReaderOp
:
public
framework
::
OperatorBase
{
...
...
@@ -104,12 +121,14 @@ class CreateDoubleBufferReaderOpMaker : public DecoratedReaderMakerBase {
};
void
DoubleBufferReader
::
ReadNext
(
std
::
vector
<
framework
::
LoDTensor
>*
out
)
{
out
->
clear
();
if
(
local_buffer_
.
empty
())
{
buffer_
->
Receive
(
out
);
}
else
{
*
out
=
local_buffer_
;
local_buffer_
.
clear
();
if
(
local_buffer_
.
payloads_
.
empty
())
{
buffer_
->
Receive
(
&
local_buffer_
);
}
*
out
=
local_buffer_
.
payloads_
;
local_buffer_
.
payloads_
.
clear
();
if
(
local_buffer_
.
ctx_
)
{
local_buffer_
.
ctx_
->
Wait
();
}
}
...
...
@@ -121,16 +140,22 @@ void DoubleBufferReader::ReInit() {
void
DoubleBufferReader
::
PrefetchThreadFunc
()
{
VLOG
(
5
)
<<
"A new prefetch thread starts."
;
size_t
gpu_ctx_offset
=
0
;
while
(
reader_
->
HasNext
())
{
std
::
vector
<
framework
::
LoDTensor
>
batch
;
reader_
->
ReadNext
(
&
batch
);
Item
batch
;
reader_
->
ReadNext
(
&
batch
.
payloads_
);
if
(
platform
::
is_gpu_place
(
place_
))
{
std
::
vector
<
framework
::
LoDTensor
>
gpu_batch
;
gpu_batch
.
resize
(
batch
.
size
());
for
(
size_t
i
=
0
;
i
<
batch
.
size
();
++
i
)
{
framework
::
TensorCopy
(
batch
[
i
],
place_
,
&
gpu_batch
[
i
]);
gpu_batch
[
i
].
set_lod
(
batch
[
i
].
lod
());
auto
&
gpu_ctx
=
this
->
ctxs_
[
gpu_ctx_offset
++
];
gpu_ctx_offset
%=
this
->
ctxs_
.
size
();
gpu_batch
.
resize
(
batch
.
payloads_
.
size
());
for
(
size_t
i
=
0
;
i
<
batch
.
payloads_
.
size
();
++
i
)
{
framework
::
TensorCopy
(
batch
.
payloads_
[
i
],
place_
,
*
gpu_ctx
,
&
gpu_batch
[
i
]);
gpu_batch
[
i
].
set_lod
(
batch
.
payloads_
[
i
].
lod
());
}
batch
.
ctx_
=
gpu_ctx
.
get
();
std
::
swap
(
gpu_batch
,
batch
.
payloads_
);
}
if
(
!
buffer_
->
Send
(
&
batch
))
{
...
...
@@ -143,7 +168,7 @@ void DoubleBufferReader::PrefetchThreadFunc() {
}
bool
DoubleBufferReader
::
HasNext
()
const
{
if
(
local_buffer_
.
empty
())
{
if
(
local_buffer_
.
payloads_
.
empty
())
{
bool
ok
=
buffer_
->
Receive
(
&
local_buffer_
);
return
ok
;
}
else
{
...
...
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