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
wk1234123
Paddle
提交
2d13462a
P
Paddle
项目概览
wk1234123
/
Paddle
与 Fork 源项目一致
Fork自
PaddlePaddle / Paddle
通知
4
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
2d13462a
编写于
9月 19, 2016
作者:
L
liaogang
浏览文件
操作
浏览文件
下载
电子邮件补丁
差异文件
Fix incompatible on CUDA atomicAdd operation
上级
4e37b226
变更
7
隐藏空白更改
内联
并排
Showing
7 changed file
with
44 addition
and
35 deletion
+44
-35
paddle/cuda/include/hl_device_functions.cuh
paddle/cuda/include/hl_device_functions.cuh
+29
-20
paddle/cuda/include/hl_gpu_lstm.cuh
paddle/cuda/include/hl_gpu_lstm.cuh
+3
-3
paddle/cuda/src/hl_cuda_lstm.cu
paddle/cuda/src/hl_cuda_lstm.cu
+3
-3
paddle/cuda/src/hl_cuda_matrix.cu
paddle/cuda/src/hl_cuda_matrix.cu
+2
-2
paddle/cuda/src/hl_cuda_sequence.cu
paddle/cuda/src/hl_cuda_sequence.cu
+1
-1
paddle/cuda/src/hl_cuda_sparse.cuh
paddle/cuda/src/hl_cuda_sparse.cuh
+5
-5
paddle/cuda/src/hl_table_apply.cu
paddle/cuda/src/hl_table_apply.cu
+1
-1
未找到文件。
paddle/cuda/include/hl_device_functions.cuh
浏览文件 @
2d13462a
...
...
@@ -16,28 +16,37 @@ limitations under the License. */
#ifndef HL_DEVICE_FUNCTIONS_CUH_
#define HL_DEVICE_FUNCTIONS_CUH_
namespace
hppl
{
static
__inline__
__device__
double
atomicAdd
(
double
*
address
,
double
val
)
{
// NOLINTNEXTLINE
unsigned
long
long
int
*
address_as_ull
=
(
unsigned
long
long
int
*
)
address
;
unsigned
long
long
int
old
=
*
address_as_ull
,
assumed
;
// NOLINT
do
{
assumed
=
old
;
old
=
atomicCAS
(
address_as_ull
,
assumed
,
__double_as_longlong
(
val
+
__longlong_as_double
(
assumed
)));
}
while
(
assumed
!=
old
);
return
__longlong_as_double
(
old
);
}
namespace
paddle
{
template
<
class
T
>
inline
__device__
T
paddleAtomicAdd
(
T
*
address
,
T
val
);
}
// namespace hppl
template
<
>
inline
__device__
float
paddleAtomicAdd
(
float
*
address
,
float
val
)
{
return
atomicAdd
(
address
,
val
);
}
#if defined(__CUDA_ARCH__) && __CUDA_ARCH__ < 600
using
hppl
::
atomicAdd
;
template
<
>
inline
__device__
double
paddleAtomicAdd
(
double
*
address
,
double
val
)
{
#if defined(__CUDA_ARCH__) && __CUDA_ARCH__ >= 600
return
atomicAdd
(
address
,
val
);
#else
// NOLINTNEXTLINE
unsigned
long
long
int
*
address_as_ull
=
(
unsigned
long
long
int
*
)
address
;
unsigned
long
long
int
old
=
*
address_as_ull
,
assumed
;
// NOLINT
do
{
assumed
=
old
;
old
=
atomicCAS
(
address_as_ull
,
assumed
,
__double_as_longlong
(
val
+
__longlong_as_double
(
assumed
)));
}
while
(
assumed
!=
old
);
return
__longlong_as_double
(
old
);
#endif
}
}
// namespace paddle
#endif
/* HL_DEVICE_FUNCTIONS_CUH_ */
paddle/cuda/include/hl_gpu_lstm.cuh
浏览文件 @
2d13462a
...
...
@@ -192,10 +192,10 @@ __global__ void KeLstmBackward(Op op,
if
(
isBatch
)
{
if
(
value
.
prevStateValue
)
{
if
(
grad
.
checkIgGrad
)
a
tomicAdd
(
grad
.
checkIgGrad
+
frameIdx
,
rCheckIGrad
);
if
(
grad
.
checkFgGrad
)
a
tomicAdd
(
grad
.
checkFgGrad
+
frameIdx
,
rCheckFGrad
);
if
(
grad
.
checkIgGrad
)
paddle
::
paddleA
tomicAdd
(
grad
.
checkIgGrad
+
frameIdx
,
rCheckIGrad
);
if
(
grad
.
checkFgGrad
)
paddle
::
paddleA
tomicAdd
(
grad
.
checkFgGrad
+
frameIdx
,
rCheckFGrad
);
}
if
(
grad
.
checkOgGrad
)
a
tomicAdd
(
grad
.
checkOgGrad
+
frameIdx
,
rCheckOGrad
);
if
(
grad
.
checkOgGrad
)
paddle
::
paddleA
tomicAdd
(
grad
.
checkOgGrad
+
frameIdx
,
rCheckOGrad
);
}
else
{
if
(
value
.
prevStateValue
)
{
if
(
grad
.
checkIgGrad
)
grad
.
checkIgGrad
[
frameIdx
]
+=
rCheckIGrad
;
...
...
paddle/cuda/src/hl_cuda_lstm.cu
浏览文件 @
2d13462a
...
...
@@ -564,11 +564,11 @@ __global__ void KeLstmBackward(real *gateValue,
/* TODO: Temporary save & merger in another kernel */
if
(
frameIdy
==
1
)
{
if
(
checkIgGrad
)
a
tomicAdd
(
checkIgGrad
+
frameIdx
,
rCheckGrad
);
if
(
checkIgGrad
)
paddle
::
paddleA
tomicAdd
(
checkIgGrad
+
frameIdx
,
rCheckGrad
);
}
else
if
(
frameIdy
==
2
)
{
if
(
checkFgGrad
)
a
tomicAdd
(
checkFgGrad
+
frameIdx
,
rCheckGrad
);
if
(
checkFgGrad
)
paddle
::
paddleA
tomicAdd
(
checkFgGrad
+
frameIdx
,
rCheckGrad
);
}
else
if
(
frameIdy
==
3
)
{
if
(
checkOgGrad
)
a
tomicAdd
(
checkOgGrad
+
frameIdx
,
rCheckGrad
);
if
(
checkOgGrad
)
paddle
::
paddleA
tomicAdd
(
checkOgGrad
+
frameIdx
,
rCheckGrad
);
}
}
...
...
paddle/cuda/src/hl_cuda_matrix.cu
浏览文件 @
2d13462a
...
...
@@ -623,7 +623,7 @@ __global__ void KeCosSimDerivative(real* grad,
prevGradY
[
index
]
+=
scale
*
grad
[
ty
]
*
prevOutX
[
index
]
*
reciprocal
;
}
else
{
a
tomicAdd
(
prevGradY
+
index
,
paddle
::
paddleA
tomicAdd
(
prevGradY
+
index
,
scale
*
grad
[
ty
]
*
prevOutX
[
index
]
*
reciprocal
);
}
}
...
...
@@ -640,7 +640,7 @@ __global__ void KeCosSimDerivative(real* grad,
(
prevOutX
[
index
]
*
reciprocalXY
-
prevOutY
[
index
]
*
reciprocalSquareSumY
);
}
else
{
a
tomicAdd
(
prevGradY
+
index
,
output
[
ty
]
*
grad
[
ty
]
*
paddle
::
paddleA
tomicAdd
(
prevGradY
+
index
,
output
[
ty
]
*
grad
[
ty
]
*
(
prevOutX
[
index
]
*
reciprocalXY
-
prevOutY
[
index
]
*
reciprocalSquareSumY
));
}
...
...
paddle/cuda/src/hl_cuda_sequence.cu
浏览文件 @
2d13462a
...
...
@@ -362,7 +362,7 @@ __global__ void KeMatrixAddRows(real* output,
if
(
AddRow
==
0
)
{
outputData
[
i
]
+=
tableData
[
i
];
}
else
{
a
tomicAdd
(
&
tableData
[
i
],
outputData
[
i
]);
paddle
::
paddleA
tomicAdd
(
&
tableData
[
i
],
outputData
[
i
]);
}
}
}
...
...
paddle/cuda/src/hl_cuda_sparse.cuh
浏览文件 @
2d13462a
...
...
@@ -280,7 +280,7 @@ __global__ void KeSMatrixCscMulDense(real *C_d,
if
(
index_n_t
<
dimN
)
{
real
tmp
;
tmp
=
alpha
*
a_r
*
b_r
[
n
];
a
tomicAdd
(
C_d_r
,
tmp
);
paddle
::
paddleA
tomicAdd
(
C_d_r
,
tmp
);
C_d_r
+=
CU_CSC_MUL_DENSE_THREAD_X
;
index_n_t
+=
CU_CSC_MUL_DENSE_THREAD_X
;
}
...
...
@@ -328,7 +328,7 @@ __global__ void KeSMatrixCscMulDense(real *C_d,
if
(
index_n_t
<
dimN
)
{
real
tmp
;
tmp
=
alpha
*
a_r
*
b_r
[
n
];
a
tomicAdd
(
C_d_r
,
tmp
);
paddle
::
paddleA
tomicAdd
(
C_d_r
,
tmp
);
C_d_r
+=
CU_CSC_MUL_DENSE_THREAD_X
;
index_n_t
+=
CU_CSC_MUL_DENSE_THREAD_X
;
}
...
...
@@ -629,7 +629,7 @@ __global__ void KeSMatrixDenseMulCsr(real *C_d,
for
(
int
n
=
0
;
n
<
CU_DM_CSR_N
;
n
++
)
{
if
(
index_m_t
++
<
dimM
)
{
tmp
=
alpha
*
b_r
*
a_r
[
n
];
a
tomicAdd
(
C_d_r
,
tmp
);
paddle
::
paddleA
tomicAdd
(
C_d_r
,
tmp
);
C_d_r
+=
dimN
;
}
}
...
...
@@ -660,7 +660,7 @@ __global__ void KeSMatrixDenseMulCsr(real *C_d,
for
(
int
n
=
0
;
n
<
CU_DM_CSR_N
;
n
++
)
{
if
(
index_m_t
++
<
dimM
)
{
tmp
=
alpha
*
b_r
*
a_r
[
n
];
a
tomicAdd
(
C_d_r
,
tmp
);
paddle
::
paddleA
tomicAdd
(
C_d_r
,
tmp
);
C_d_r
+=
dimN
;
}
}
...
...
@@ -912,7 +912,7 @@ __global__ void KeSMatrixCsrColumnSum(real* a_val, real* csr_val,
for
(
int
idx
=
gid
;
idx
<
dimNNZ
;
idx
+=
gridDim
.
x
*
blockDim
.
x
)
{
int
colIdx
=
csr_col
[
idx
];
real
val
=
csr_val
[
idx
];
a
tomicAdd
(
a_val
+
colIdx
,
val
);
paddle
::
paddleA
tomicAdd
(
a_val
+
colIdx
,
val
);
}
}
...
...
paddle/cuda/src/hl_table_apply.cu
浏览文件 @
2d13462a
...
...
@@ -35,7 +35,7 @@ __global__ void KeMatrixAddRows(real* output, int ldo,
real
*
tab
=
table
+
tableId
*
ldt
;
for
(
int
i
=
idx
;
i
<
dim
;
i
+=
blockDimX
)
{
if
(
AddRow
)
{
a
tomicAdd
(
&
tab
[
i
],
out
[
i
]);
paddle
::
paddleA
tomicAdd
(
&
tab
[
i
],
out
[
i
]);
}
else
{
out
[
i
]
+=
tab
[
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