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
wmsofts
Paddle
提交
303fb789
P
Paddle
项目概览
wmsofts
/
Paddle
与 Fork 源项目一致
Fork自
PaddlePaddle / Paddle
通知
1
Star
0
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
体验新版 GitCode,发现更多精彩内容 >>
提交
303fb789
编写于
7月 28, 2017
作者:
Q
qijun
浏览文件
操作
浏览文件
下载
电子邮件补丁
差异文件
refine tensor copy from
上级
5364b394
变更
4
隐藏空白更改
内联
并排
Showing
4 changed file
with
15 addition
and
36 deletion
+15
-36
paddle/framework/detail/tensor-inl.h
paddle/framework/detail/tensor-inl.h
+10
-24
paddle/framework/tensor.h
paddle/framework/tensor.h
+1
-8
paddle/memory/memcpy.cc
paddle/memory/memcpy.cc
+3
-3
paddle/memory/memcpy.h
paddle/memory/memcpy.h
+1
-1
未找到文件。
paddle/framework/detail/tensor-inl.h
浏览文件 @
303fb789
...
...
@@ -83,7 +83,7 @@ inline void Tensor::ShareDataWith(const Tensor& src) {
template
<
typename
T
>
inline
void
Tensor
::
CopyFrom
(
const
Tensor
&
src
,
const
platform
::
CPU
Place
&
dst_place
)
{
const
platform
::
Place
&
dst_place
)
{
src
.
check_memory_size
<
T
>
();
Resize
(
src
.
dims
());
...
...
@@ -94,41 +94,27 @@ inline void Tensor::CopyFrom(const Tensor& src,
auto
size
=
product
(
src
.
dims_
)
*
sizeof
(
T
);
if
(
platform
::
is_cpu_place
(
src_place
))
{
if
(
platform
::
is_cpu_place
(
src_place
)
&&
platform
::
is_cpu_place
(
dst_place
)
)
{
memory
::
Copy
(
boost
::
get
<
platform
::
CPUPlace
>
(
dst_place
),
dst_ptr
,
boost
::
get
<
platform
::
CPUPlace
>
(
src_place
),
src_ptr
,
size
);
}
#ifndef PADDLE_ONLY_CPU
else
if
(
platform
::
is_gpu_place
(
src_place
))
{
else
if
(
platform
::
is_gpu_place
(
src_place
)
&&
platform
::
is_cpu_place
(
dst_place
))
{
memory
::
Copy
(
boost
::
get
<
platform
::
CPUPlace
>
(
dst_place
),
dst_ptr
,
boost
::
get
<
platform
::
GPUPlace
>
(
src_place
),
src_ptr
,
size
,
0
);
}
#endif
}
#ifndef PADDLE_ONLY_CPU
template
<
typename
T
>
inline
void
Tensor
::
CopyFrom
(
const
Tensor
&
src
,
const
platform
::
GPUPlace
&
dst_place
)
{
src
.
check_memory_size
<
T
>
();
Resize
(
src
.
dims
());
auto
src_place
=
src
.
holder_
->
place
();
auto
src_ptr
=
static_cast
<
const
void
*>
(
src
.
data
<
T
>
());
auto
dst_ptr
=
static_cast
<
void
*>
(
mutable_data
<
T
>
(
dst_place
));
auto
size
=
product
(
src
.
dims_
)
*
sizeof
(
T
);
if
(
platform
::
is_cpu_place
(
src_place
))
{
}
else
if
(
platform
::
is_cpu_place
(
src_place
)
&&
platform
::
is_gpu_place
(
dst_place
))
{
memory
::
Copy
(
boost
::
get
<
platform
::
GPUPlace
>
(
dst_place
),
dst_ptr
,
boost
::
get
<
platform
::
CPUPlace
>
(
src_place
),
src_ptr
,
size
,
0
);
}
else
if
(
platform
::
is_gpu_place
(
src_place
))
{
}
else
if
(
platform
::
is_gpu_place
(
src_place
)
&&
platform
::
is_gpu_place
(
dst_place
))
{
memory
::
Copy
(
boost
::
get
<
platform
::
GPUPlace
>
(
dst_place
),
dst_ptr
,
boost
::
get
<
platform
::
GPUPlace
>
(
src_place
),
src_ptr
,
size
,
0
);
}
}
#endif
}
template
<
typename
T
>
inline
Tensor
Tensor
::
Slice
(
const
int
&
begin_idx
,
const
int
&
end_idx
)
const
{
...
...
paddle/framework/tensor.h
浏览文件 @
303fb789
...
...
@@ -94,14 +94,7 @@ class Tensor {
* @note CopyFrom supports CPU <-> GPU, GPU <-> GPU.
*/
template
<
typename
T
>
inline
void
CopyFrom
(
const
Tensor
&
src
,
const
platform
::
CPUDeviceContext
&
ctx
);
#ifndef PADDLE_ONLY_CPU
template
<
typename
T
>
inline
void
CopyFrom
(
const
Tensor
&
src
,
const
platform
::
CUDADeviceContext
&
ctx
);
#endif
inline
void
CopyFrom
(
const
Tensor
&
src
,
const
platform
::
Place
&
dst_place
);
/**
* @brief Return the slice of the tensor.
...
...
paddle/memory/memcpy.cc
浏览文件 @
303fb789
...
...
@@ -34,7 +34,7 @@ void Copy<platform::CPUPlace, platform::GPUPlace>(platform::CPUPlace dst_place,
void
*
dst
,
platform
::
GPUPlace
src_place
,
const
void
*
src
,
size_t
num
,
cudaStream_t
stream
=
0
)
{
cudaStream_t
stream
)
{
platform
::
SetDeviceId
(
src_place
.
device
);
platform
::
GpuMemcpyAsync
(
dst
,
src
,
num
,
cudaMemcpyDeviceToHost
,
stream
);
}
...
...
@@ -44,7 +44,7 @@ void Copy<platform::GPUPlace, platform::CPUPlace>(platform::GPUPlace dst_place,
void
*
dst
,
platform
::
CPUPlace
src_place
,
const
void
*
src
,
size_t
num
,
cudaStream_t
stream
=
0
)
{
cudaStream_t
stream
)
{
platform
::
SetDeviceId
(
dst_place
.
device
);
platform
::
GpuMemcpyAsync
(
dst
,
src
,
num
,
cudaMemcpyHostToDevice
,
stream
);
}
...
...
@@ -54,7 +54,7 @@ void Copy<platform::GPUPlace, platform::GPUPlace>(platform::GPUPlace dst_place,
void
*
dst
,
platform
::
GPUPlace
src_place
,
const
void
*
src
,
size_t
num
,
cudaStream_t
stream
=
0
)
{
cudaStream_t
stream
)
{
if
(
dst_place
==
src_place
)
{
platform
::
SetDeviceId
(
src_place
.
device
);
platform
::
GpuMemcpyAsync
(
dst
,
src
,
num
,
cudaMemcpyDeviceToDevice
,
stream
);
...
...
paddle/memory/memcpy.h
浏览文件 @
303fb789
...
...
@@ -51,7 +51,7 @@ void Copy(DstPlace, void* dst, SrcPlace, const void* src, size_t num);
*/
template
<
typename
DstPlace
,
typename
SrcPlace
>
void
Copy
(
DstPlace
,
void
*
dst
,
SrcPlace
,
const
void
*
src
,
size_t
num
,
cudaStream_t
stream
=
0
);
cudaStream_t
stream
);
#endif // PADDLE_ONLY_CPU
...
...
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