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
384372f5
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
未验证
提交
384372f5
编写于
7月 27, 2021
作者:
W
wuhuachaocoding
提交者:
GitHub
7月 27, 2021
浏览文件
操作
浏览文件
下载
电子邮件补丁
差异文件
fix concat bug (#34319) (#34396)
上级
862e81ef
变更
2
显示空白变更内容
内联
并排
Showing
2 changed file
with
111 addition
and
48 deletion
+111
-48
paddle/fluid/operators/math/concat_and_split.cc
paddle/fluid/operators/math/concat_and_split.cc
+9
-9
paddle/fluid/operators/math/concat_and_split.cu
paddle/fluid/operators/math/concat_and_split.cu
+102
-39
未找到文件。
paddle/fluid/operators/math/concat_and_split.cc
浏览文件 @
384372f5
...
...
@@ -40,18 +40,18 @@ class ConcatFunctor<platform::CPUDeviceContext, T> {
const
std
::
vector
<
framework
::
Tensor
>&
input
,
int
axis
,
framework
::
Tensor
*
output
)
{
// TODO(zcd): Add input data validity checking
in
t
num
=
input
.
size
();
size_
t
num
=
input
.
size
();
int
rows
=
1
;
int
64_t
rows
=
1
;
auto
dim_0
=
input
[
0
].
dims
();
for
(
int
i
=
0
;
i
<
axis
;
++
i
)
{
rows
*=
dim_0
[
i
];
}
int
out_rows
=
rows
,
out_cols
=
0
;
int
64_t
out_rows
=
rows
,
out_cols
=
0
;
std
::
vector
<
int64_t
>
input_cols
(
input
.
size
());
for
(
in
t
i
=
0
;
i
<
num
;
++
i
)
{
int
t_cols
=
input
[
i
].
numel
()
/
rows
;
for
(
size_
t
i
=
0
;
i
<
num
;
++
i
)
{
int
64_t
t_cols
=
input
[
i
].
numel
()
/
rows
;
out_cols
+=
t_cols
;
input_cols
[
i
]
=
t_cols
;
}
...
...
@@ -59,11 +59,11 @@ class ConcatFunctor<platform::CPUDeviceContext, T> {
// computation
auto
output_data
=
output
->
data
<
T
>
();
int
col_idx
=
0
;
for
(
in
t
j
=
0
;
j
<
num
;
++
j
)
{
int
col_len
=
input_cols
[
j
];
int
64_t
col_idx
=
0
;
for
(
size_
t
j
=
0
;
j
<
num
;
++
j
)
{
int
64_t
col_len
=
input_cols
[
j
];
auto
input_data
=
input
[
j
].
data
<
T
>
();
for
(
int
k
=
0
;
k
<
out_rows
;
++
k
)
{
for
(
int
64_t
k
=
0
;
k
<
out_rows
;
++
k
)
{
memory
::
Copy
(
cpu_place
,
output_data
+
k
*
out_cols
+
col_idx
,
cpu_place
,
input_data
+
k
*
col_len
,
sizeof
(
T
)
*
col_len
);
}
...
...
paddle/fluid/operators/math/concat_and_split.cu
浏览文件 @
384372f5
...
...
@@ -14,6 +14,7 @@ limitations under the License. */
#include <algorithm>
#include <vector>
#include "gflags/gflags.h"
#include "paddle/fluid/framework/mixed_vector.h"
#include "paddle/fluid/memory/malloc.h"
#include "paddle/fluid/operators/math/concat_and_split.h"
...
...
@@ -25,9 +26,9 @@ namespace operators {
namespace
math
{
template
<
typename
T
>
__global__
void
ConcatKernel
(
const
T
**
inputs
,
const
int
*
input_cols
,
int
col_size
,
const
int
output_rows
,
const
int
output_cols
,
T
*
output
)
{
__global__
void
ConcatKernel
(
const
T
**
inputs
,
const
int
64_t
*
input_cols
,
int
col_size
,
const
int
64_t
output_rows
,
const
int
64_t
output_cols
,
T
*
output
)
{
int
tid_x
=
blockIdx
.
x
*
blockDim
.
x
+
threadIdx
.
x
;
int
curr_segment
=
0
;
int
curr_offset
=
input_cols
[
0
];
...
...
@@ -69,8 +70,8 @@ __device__ void ConcatKernelDetail(const T** inputs_data,
template
<
typename
T
>
__global__
void
ConcatKernel
(
const
T
*
input_addr0
,
const
T
*
input_addr1
,
const
int
fixed_in_col
,
const
in
t
out_rows
,
const
int
out_cols
,
T
*
output_data
)
{
const
int
64_t
fixed_in_col
,
const
int64_
t
out_rows
,
const
int
64_t
out_cols
,
T
*
output_data
)
{
const
T
*
inputs_data
[
2
];
inputs_data
[
0
]
=
input_addr0
;
inputs_data
[
1
]
=
input_addr1
;
...
...
@@ -80,8 +81,8 @@ __global__ void ConcatKernel(const T* input_addr0, const T* input_addr1,
template
<
typename
T
>
__global__
void
ConcatKernel
(
const
T
*
input_addr0
,
const
T
*
input_addr1
,
const
T
*
input_addr2
,
const
int
fixed_in_col
,
const
int
out_rows
,
const
in
t
out_cols
,
const
T
*
input_addr2
,
const
int
64_t
fixed_in_col
,
const
int
64_t
out_rows
,
const
int64_
t
out_cols
,
T
*
output_data
)
{
const
T
*
inputs_data
[
3
];
inputs_data
[
0
]
=
input_addr0
;
...
...
@@ -94,8 +95,8 @@ __global__ void ConcatKernel(const T* input_addr0, const T* input_addr1,
template
<
typename
T
>
__global__
void
ConcatKernel
(
const
T
*
input_addr0
,
const
T
*
input_addr1
,
const
T
*
input_addr2
,
const
T
*
input_addr3
,
const
int
fixed_in_col
,
const
in
t
out_rows
,
const
int
out_cols
,
T
*
output_data
)
{
const
int
64_t
fixed_in_col
,
const
int64_
t
out_rows
,
const
int
64_t
out_cols
,
T
*
output_data
)
{
const
T
*
inputs_data
[
4
];
inputs_data
[
0
]
=
input_addr0
;
inputs_data
[
1
]
=
input_addr1
;
...
...
@@ -107,8 +108,8 @@ __global__ void ConcatKernel(const T* input_addr0, const T* input_addr1,
template
<
typename
T
>
__global__
void
ConcatKernel
(
const
T
**
inputs_data
,
const
int
in_num
,
const
int
fixed_in_col
,
const
in
t
out_rows
,
const
int
out_cols
,
T
*
output_data
)
{
const
int
64_t
fixed_in_col
,
const
int64_
t
out_rows
,
const
int
64_t
out_cols
,
T
*
output_data
)
{
ConcatKernelDetail
<
T
>
(
inputs_data
,
fixed_in_col
,
out_rows
,
out_cols
,
output_data
);
}
...
...
@@ -234,21 +235,41 @@ class ConcatFunctor<platform::CUDADeviceContext, T> {
framework
::
Tensor
*
output
)
{
// TODO(zcd): Add input data validity checking
int
in_num
=
input
.
size
();
int
in_row
=
1
;
int
64_t
in_row
=
1
;
auto
dim_0
=
input
[
0
].
dims
();
for
(
int
i
=
0
;
i
<
axis
;
++
i
)
{
in_row
*=
dim_0
[
i
];
}
int
in_col
=
input
[
0
].
numel
()
/
in_row
;
int
out_row
=
in_row
,
out_col
=
0
;
std
::
vector
<
const
T
*>
inputs_data
(
in_num
);
std
::
vector
<
int
>
inputs_col
(
in_num
+
1
);
int64_t
in_col
=
input
[
0
].
numel
()
/
in_row
;
int64_t
out_row
=
in_row
,
out_col
=
0
;
int
inputs_col_num
=
in_num
+
1
;
std
::
vector
<
const
T
*>
inputs_data_vec
(
in_num
);
std
::
vector
<
int64_t
>
inputs_col_vec
(
inputs_col_num
);
const
T
**
inputs_data
=
inputs_data_vec
.
data
();
int64_t
*
inputs_col
=
inputs_col_vec
.
data
();
// There are some differences between hip runtime and NV runtime.
// In NV, when the pageable memory data less than 64K is transferred from
// hosttodevice, it will be automatically asynchronous.
// However, only pinned memory in hip can copy asynchronously
// https://docs.nvidia.com/cuda/cuda-c-programming-guide/index.html#concurrent-execution-host-device
// 3.2.6.1. Concurrent Execution between Host and Device
// Memory copies from host to device of a memory block of 64 KB or less
#ifdef PADDLE_WITH_HIP
memory
::
AllocationPtr
data_alloc
,
col_alloc
;
data_alloc
=
memory
::
Alloc
(
platform
::
CUDAPinnedPlace
(),
in_num
*
sizeof
(
T
*
));
inputs_data
=
reinterpret_cast
<
const
T
**>
(
data_alloc
->
ptr
());
col_alloc
=
memory
::
Alloc
(
platform
::
CUDAPinnedPlace
(),
inputs_col_num
*
sizeof
(
int
));
inputs_col
=
reinterpret_cast
<
int64_t
*>
(
col_alloc
->
ptr
());
#endif
inputs_col
[
0
]
=
0
;
bool
has_same_shape
=
true
;
for
(
int
i
=
0
;
i
<
in_num
;
++
i
)
{
int
t_cols
=
input
[
i
].
numel
()
/
in_row
;
int
64_t
t_cols
=
input
[
i
].
numel
()
/
in_row
;
if
(
has_same_shape
)
{
if
(
t_cols
!=
in_col
)
has_same_shape
=
false
;
}
...
...
@@ -264,12 +285,11 @@ class ConcatFunctor<platform::CUDADeviceContext, T> {
memory
::
allocation
::
AllocationPtr
tmp_dev_ins_data
;
const
T
**
dev_ins_data
=
nullptr
;
if
(
!
has_same_shape
||
in_num
<
2
||
in_num
>
4
)
{
tmp_dev_ins_data
=
memory
::
Alloc
(
context
,
inputs_data
.
size
()
*
sizeof
(
T
*
));
tmp_dev_ins_data
=
memory
::
Alloc
(
context
,
in_num
*
sizeof
(
T
*
));
memory
::
Copy
(
BOOST_GET_CONST
(
platform
::
CUDAPlace
,
context
.
GetPlace
()),
tmp_dev_ins_data
->
ptr
(),
platform
::
CPUPlace
(),
static_cast
<
void
*>
(
inputs_data
.
data
()
),
inputs_data
.
size
()
*
sizeof
(
T
*
),
context
.
stream
());
static_cast
<
void
*>
(
inputs_data
),
in_num
*
sizeof
(
T
*
),
context
.
stream
());
dev_ins_data
=
reinterpret_cast
<
const
T
**>
(
tmp_dev_ins_data
->
ptr
());
}
...
...
@@ -292,17 +312,31 @@ class ConcatFunctor<platform::CUDADeviceContext, T> {
}
}
else
{
auto
tmp_dev_ins_col_data
=
memory
::
Alloc
(
context
,
inputs_col
.
size
()
*
sizeof
(
in
t
));
memory
::
Alloc
(
context
,
inputs_col
_num
*
sizeof
(
int64_
t
));
memory
::
Copy
(
BOOST_GET_CONST
(
platform
::
CUDAPlace
,
context
.
GetPlace
()),
tmp_dev_ins_col_data
->
ptr
(),
platform
::
CPUPlace
(),
static_cast
<
void
*>
(
inputs_col
.
data
()),
inputs_col
.
size
()
*
sizeof
(
int
),
context
.
stream
());
int
*
dev_ins_col_data
=
static_cast
<
int
*>
(
tmp_dev_ins_col_data
->
ptr
());
static_cast
<
void
*>
(
inputs_col
),
inputs_col_num
*
sizeof
(
int64_t
),
context
.
stream
());
int64_t
*
dev_ins_col_data
=
static_cast
<
int64_t
*>
(
tmp_dev_ins_col_data
->
ptr
());
ConcatKernel
<<<
grid_dims
,
block_dims
,
0
,
context
.
stream
()
>>>
(
dev_ins_data
,
dev_ins_col_data
,
static_cast
<
int
>
(
inputs_col
.
size
()
),
dev_ins_data
,
dev_ins_col_data
,
static_cast
<
int
>
(
inputs_col
_num
),
out_row
,
out_col
,
output
->
data
<
T
>
());
}
#ifdef PADDLE_WITH_HIP
// Prevent the pinned memory value from being covered and release the memory
// after the launch kernel of the stream is executed (reapply pinned memory
// next time)
auto
*
data_alloc_released
=
data_alloc
.
release
();
auto
*
col_alloc_released
=
col_alloc
.
release
();
context
.
AddStreamCallback
([
data_alloc_released
,
col_alloc_released
]
{
memory
::
allocation
::
AllocationDeleter
deleter
;
deleter
(
data_alloc_released
);
deleter
(
col_alloc_released
);
});
#endif
}
};
...
...
@@ -313,6 +347,7 @@ class ConcatFunctor<platform::CUDADeviceContext, T> {
template
<
typename
T
>
class
SplitFunctor
<
platform
::
CUDADeviceContext
,
T
>
{
public:
SplitFunctor
();
void
operator
()(
const
platform
::
CUDADeviceContext
&
context
,
const
framework
::
Tensor
&
input
,
const
std
::
vector
<
const
framework
::
Tensor
*>&
ref_inputs
,
...
...
@@ -329,8 +364,27 @@ class SplitFunctor<platform::CUDADeviceContext, T> {
int64_t
in_col
=
0
,
in_row
=
out_row
;
bool
has_same_shape
=
true
;
std
::
vector
<
T
*>
outputs_data
(
o_num
);
std
::
vector
<
int64_t
>
outputs_cols
(
o_num
+
1
);
int
outputs_cols_num
=
o_num
+
1
;
std
::
vector
<
T
*>
outputs_data_vec
(
o_num
);
std
::
vector
<
int64_t
>
outputs_cols_vec
(
outputs_cols_num
);
T
**
outputs_data
=
outputs_data_vec
.
data
();
int64_t
*
outputs_cols
=
outputs_cols_vec
.
data
();
// There are some differences between hip runtime and NV runtime.
// In NV, when the pageable memory data less than 64K is transferred from
// hosttodevice, it will be automatically asynchronous.
// However, only pinned memory in hip can copy asynchronously
// https://docs.nvidia.com/cuda/cuda-c-programming-guide/index.html#concurrent-execution-host-device
// 3.2.6.1. Concurrent Execution between Host and Device
// Memory copies from host to device of a memory block of 64 KB or less
#ifdef PADDLE_WITH_HIP
memory
::
AllocationPtr
data_alloc
,
cols_alloc
;
data_alloc
=
memory
::
Alloc
(
platform
::
CUDAPinnedPlace
(),
o_num
*
sizeof
(
T
*
));
outputs_data
=
reinterpret_cast
<
T
**>
(
data_alloc
->
ptr
());
cols_alloc
=
memory
::
Alloc
(
platform
::
CUDAPinnedPlace
(),
(
outputs_cols_num
)
*
sizeof
(
int64_t
));
outputs_cols
=
reinterpret_cast
<
int64_t
*>
(
cols_alloc
->
ptr
());
#endif
outputs_cols
[
0
]
=
0
;
for
(
int
i
=
0
;
i
<
o_num
;
++
i
)
{
...
...
@@ -354,12 +408,11 @@ class SplitFunctor<platform::CUDADeviceContext, T> {
memory
::
allocation
::
AllocationPtr
tmp_dev_outs_data
;
T
**
dev_out_gpu_data
=
nullptr
;
if
(
!
has_same_shape
||
o_num
<
2
||
o_num
>
4
)
{
tmp_dev_outs_data
=
memory
::
Alloc
(
context
,
outputs_data
.
size
()
*
sizeof
(
T
*
));
tmp_dev_outs_data
=
memory
::
Alloc
(
context
,
o_num
*
sizeof
(
T
*
));
memory
::
Copy
(
BOOST_GET_CONST
(
platform
::
CUDAPlace
,
context
.
GetPlace
()),
tmp_dev_outs_data
->
ptr
(),
platform
::
CPUPlace
(),
reinterpret_cast
<
void
*>
(
outputs_data
.
data
()
),
outputs_data
.
size
()
*
sizeof
(
T
*
),
context
.
stream
());
reinterpret_cast
<
void
*>
(
outputs_data
),
o_num
*
sizeof
(
T
*
),
context
.
stream
());
dev_out_gpu_data
=
reinterpret_cast
<
T
**>
(
tmp_dev_outs_data
->
ptr
());
}
...
...
@@ -382,20 +435,30 @@ class SplitFunctor<platform::CUDADeviceContext, T> {
}
}
else
{
auto
tmp_dev_ins_col_data
=
memory
::
Alloc
(
context
,
outputs_cols
.
size
()
*
sizeof
(
int64_t
));
memory
::
Alloc
(
context
,
outputs_cols_num
*
sizeof
(
int64_t
));
memory
::
Copy
(
BOOST_GET_CONST
(
platform
::
CUDAPlace
,
context
.
GetPlace
()),
tmp_dev_ins_col_data
->
ptr
(),
platform
::
CPUPlace
(),
reinterpret_cast
<
void
*>
(
outputs_cols
.
data
()
),
outputs_cols
.
size
()
*
sizeof
(
int64_t
),
context
.
stream
());
reinterpret_cast
<
void
*>
(
outputs_cols
),
outputs_cols
_num
*
sizeof
(
int64_t
),
context
.
stream
());
int64_t
*
dev_outs_col_data
=
reinterpret_cast
<
int64_t
*>
(
tmp_dev_ins_col_data
->
ptr
());
SplitKernel
<<<
grid_dims
,
block_dims
,
0
,
context
.
stream
()
>>>
(
input
.
data
<
T
>
(),
in_row
,
in_col
,
dev_outs_col_data
,
static_cast
<
int
>
(
outputs_cols
.
size
()
),
dev_out_gpu_data
);
static_cast
<
int
>
(
outputs_cols
_num
),
dev_out_gpu_data
);
}
#ifdef PADDLE_WITH_HIP
// Prevent the pinned memory value from being covered and release the memory
// after the launch kernel of the stream is executed (reapply pinned memory
// next time)
auto
*
data_alloc_released
=
data_alloc
.
release
();
auto
*
cols_alloc_released
=
cols_alloc
.
release
();
context
.
AddStreamCallback
([
data_alloc_released
,
cols_alloc_released
]
{
memory
::
allocation
::
AllocationDeleter
deleter
;
deleter
(
data_alloc_released
);
deleter
(
cols_alloc_released
);
});
#endif
}
};
...
...
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