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
sven357
Paddle
提交
494c262a
P
Paddle
项目概览
sven357
/
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
494c262a
编写于
4月 14, 2018
作者:
W
whs
提交者:
Yi Wang
4月 13, 2018
浏览文件
操作
浏览文件
下载
电子邮件补丁
差异文件
Fix average_accumulate_op for parallel executor. (#9852)
上级
630943c7
变更
1
显示空白变更内容
内联
并排
Showing
1 changed file
with
12 addition
and
8 deletion
+12
-8
paddle/fluid/operators/average_accumulates_op.cu
paddle/fluid/operators/average_accumulates_op.cu
+12
-8
未找到文件。
paddle/fluid/operators/average_accumulates_op.cu
浏览文件 @
494c262a
...
...
@@ -25,12 +25,14 @@ void GetAccumulators<paddle::platform::CUDADeviceContext>(
auto
*
in_num_accumulates
=
ctx
.
Input
<
Tensor
>
(
"in_num_accumulates"
);
auto
*
in_num_updates
=
ctx
.
Input
<
Tensor
>
(
"in_num_updates"
);
auto
stream
=
ctx
.
cuda_device_context
().
stream
();
memory
::
Copy
(
platform
::
CPUPlace
(),
old_num_accumulates_
,
platform
::
CUDAPlace
(),
in_old_num_accumulates
->
data
<
int64_t
>
(),
sizeof
(
int64_t
),
stream
);
memory
::
Copy
(
platform
::
CPUPlace
(),
num_accumulates_
,
platform
::
CUDAPlace
(),
auto
cuda_place
=
boost
::
get
<
platform
::
CUDAPlace
>
(
in_old_num_accumulates
->
place
());
memory
::
Copy
(
platform
::
CPUPlace
(),
old_num_accumulates_
,
cuda_place
,
in_old_num_accumulates
->
data
<
int64_t
>
(),
sizeof
(
int64_t
),
stream
);
memory
::
Copy
(
platform
::
CPUPlace
(),
num_accumulates_
,
cuda_place
,
in_num_accumulates
->
data
<
int64_t
>
(),
sizeof
(
int64_t
),
stream
);
memory
::
Copy
(
platform
::
CPUPlace
(),
num_updates_
,
platform
::
CUDAPlace
()
,
memory
::
Copy
(
platform
::
CPUPlace
(),
num_updates_
,
cuda_place
,
in_num_updates
->
data
<
int64_t
>
(),
sizeof
(
int64_t
),
stream
);
}
...
...
@@ -42,14 +44,16 @@ void SetAccumulators<paddle::platform::CUDADeviceContext>(
auto
*
out_old_num_accumulates
=
ctx
.
Output
<
Tensor
>
(
"out_old_num_accumulates"
);
auto
*
out_num_accumulates
=
ctx
.
Output
<
Tensor
>
(
"out_num_accumulates"
);
auto
*
out_num_updates
=
ctx
.
Output
<
Tensor
>
(
"out_num_updates"
);
auto
cuda_place
=
boost
::
get
<
platform
::
CUDAPlace
>
(
out_old_num_accumulates
->
place
());
memory
::
Copy
(
platform
::
CUDAPlace
()
,
out_old_num_accumulates
->
data
<
int64_t
>
(),
memory
::
Copy
(
cuda_place
,
out_old_num_accumulates
->
data
<
int64_t
>
(),
platform
::
CPUPlace
(),
&
old_num_accumulates_
,
sizeof
(
int64_t
),
stream
);
memory
::
Copy
(
platform
::
CUDAPlace
()
,
out_num_accumulates
->
data
<
int64_t
>
(),
memory
::
Copy
(
cuda_place
,
out_num_accumulates
->
data
<
int64_t
>
(),
platform
::
CPUPlace
(),
&
num_accumulates_
,
sizeof
(
int64_t
),
stream
);
memory
::
Copy
(
platform
::
CUDAPlace
()
,
out_num_updates
->
data
<
int64_t
>
(),
memory
::
Copy
(
cuda_place
,
out_num_updates
->
data
<
int64_t
>
(),
platform
::
CPUPlace
(),
&
num_updates_
,
sizeof
(
int64_t
),
stream
);
}
...
...
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