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
wzpei9
Paddle
提交
960a5255
P
Paddle
项目概览
wzpei9
/
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
960a5255
编写于
8月 14, 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
fix gpu build error
上级
2ec8dab4
变更
4
隐藏空白更改
内联
并排
Showing
4 changed file
with
42 addition
and
79 deletion
+42
-79
paddle/operators/math/CMakeLists.txt
paddle/operators/math/CMakeLists.txt
+1
-1
paddle/operators/math/math_function.cc
paddle/operators/math/math_function.cc
+0
-38
paddle/operators/math/math_function.cu
paddle/operators/math/math_function.cu
+0
-38
paddle/operators/math/math_function.h
paddle/operators/math/math_function.h
+41
-2
未找到文件。
paddle/operators/math/CMakeLists.txt
浏览文件 @
960a5255
...
@@ -7,7 +7,7 @@ endif()
...
@@ -7,7 +7,7 @@ endif()
if
(
WITH_GPU
)
if
(
WITH_GPU
)
nv_library
(
math_function SRCS math_function.cc math_function.cu DEPS
${
BLAS_LIB
}
device_context
)
nv_library
(
math_function SRCS math_function.cc math_function.cu DEPS
${
BLAS_LIB
}
device_context
)
else
()
else
()
cc_library
(
math_function SRCS math_function.cc
math_function.cu
DEPS
${
BLAS_LIB
}
device_context
)
cc_library
(
math_function SRCS math_function.cc DEPS
${
BLAS_LIB
}
device_context
)
endif
()
endif
()
nv_test
(
math_function_test SRCS math_function_test.cc DEPS math_function tensor
)
nv_test
(
math_function_test SRCS math_function_test.cc DEPS math_function tensor
)
paddle/operators/math/math_function.cc
浏览文件 @
960a5255
...
@@ -12,44 +12,6 @@ WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
...
@@ -12,44 +12,6 @@ WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
See the License for the specific language governing permissions and
See the License for the specific language governing permissions and
limitations under the License. */
limitations under the License. */
#ifdef PADDLE_USE_MKLML
#include <mkl_cblas.h>
#include <mkl_lapacke.h>
#include <mkl_vml_functions.h>
#endif
#ifdef PADDLE_USE_MKL
#include <mkl.h>
#include <mkl_lapacke.h>
#endif
#ifdef PADDLE_USE_ATLAS
extern
"C"
{
#include <cblas.h>
#include <clapack.h>
}
#endif
#ifdef PADDLE_USE_OPENBLAS
#include <cblas.h>
#include <lapacke.h>
#endif
#ifndef LAPACK_FOUND
extern
"C"
{
#include <cblas.h>
int
LAPACKE_sgetrf
(
int
matrix_layout
,
int
m
,
int
n
,
float
*
a
,
int
lda
,
int
*
ipiv
);
int
LAPACKE_dgetrf
(
int
matrix_layout
,
int
m
,
int
n
,
double
*
a
,
int
lda
,
int
*
ipiv
);
int
LAPACKE_sgetri
(
int
matrix_layout
,
int
n
,
float
*
a
,
int
lda
,
const
int
*
ipiv
);
int
LAPACKE_dgetri
(
int
matrix_layout
,
int
n
,
double
*
a
,
int
lda
,
const
int
*
ipiv
);
}
#endif
#include <cmath>
#include "paddle/operators/math/math_function.h"
#include "paddle/operators/math/math_function.h"
namespace
paddle
{
namespace
paddle
{
...
...
paddle/operators/math/math_function.cu
浏览文件 @
960a5255
...
@@ -12,44 +12,6 @@ WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
...
@@ -12,44 +12,6 @@ WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
See the License for the specific language governing permissions and
See the License for the specific language governing permissions and
limitations under the License. */
limitations under the License. */
#ifdef PADDLE_USE_MKLML
#include <mkl_cblas.h>
#include <mkl_lapacke.h>
#include <mkl_vml_functions.h>
#endif
#ifdef PADDLE_USE_MKL
#include <mkl.h>
#include <mkl_lapacke.h>
#endif
#ifdef PADDLE_USE_ATLAS
extern
"C"
{
#include <cblas.h>
#include <clapack.h>
}
#endif
#ifdef PADDLE_USE_OPENBLAS
#include <cblas.h>
#include <lapacke.h>
#endif
#ifndef LAPACK_FOUND
extern
"C"
{
#include <cblas.h>
int
LAPACKE_sgetrf
(
int
matrix_layout
,
int
m
,
int
n
,
float
*
a
,
int
lda
,
int
*
ipiv
);
int
LAPACKE_dgetrf
(
int
matrix_layout
,
int
m
,
int
n
,
double
*
a
,
int
lda
,
int
*
ipiv
);
int
LAPACKE_sgetri
(
int
matrix_layout
,
int
n
,
float
*
a
,
int
lda
,
const
int
*
ipiv
);
int
LAPACKE_dgetri
(
int
matrix_layout
,
int
n
,
double
*
a
,
int
lda
,
const
int
*
ipiv
);
}
#endif
#include <cmath>
#include "paddle/operators/math/math_function.h"
#include "paddle/operators/math/math_function.h"
namespace
paddle
{
namespace
paddle
{
...
...
paddle/operators/math/math_function.h
浏览文件 @
960a5255
...
@@ -13,6 +13,44 @@ See the License for the specific language governing permissions and
...
@@ -13,6 +13,44 @@ See the License for the specific language governing permissions and
limitations under the License. */
limitations under the License. */
#pragma once
#pragma once
#ifdef PADDLE_USE_MKLML
#include <mkl_cblas.h>
#include <mkl_lapacke.h>
#include <mkl_vml_functions.h>
#endif
#ifdef PADDLE_USE_MKL
#include <mkl.h>
#include <mkl_lapacke.h>
#endif
#ifdef PADDLE_USE_ATLAS
extern
"C"
{
#include <cblas.h>
#include <clapack.h>
}
#endif
#ifdef PADDLE_USE_OPENBLAS
#include <cblas.h>
#include <lapacke.h>
#endif
#ifndef LAPACK_FOUND
extern
"C"
{
#include <cblas.h>
int
LAPACKE_sgetrf
(
int
matrix_layout
,
int
m
,
int
n
,
float
*
a
,
int
lda
,
int
*
ipiv
);
int
LAPACKE_dgetrf
(
int
matrix_layout
,
int
m
,
int
n
,
double
*
a
,
int
lda
,
int
*
ipiv
);
int
LAPACKE_sgetri
(
int
matrix_layout
,
int
n
,
float
*
a
,
int
lda
,
const
int
*
ipiv
);
int
LAPACKE_dgetri
(
int
matrix_layout
,
int
n
,
double
*
a
,
int
lda
,
const
int
*
ipiv
);
}
#endif
#include <cmath>
#include "paddle/framework/tensor.h"
#include "paddle/framework/tensor.h"
#include "paddle/platform/device_context.h"
#include "paddle/platform/device_context.h"
...
@@ -27,6 +65,7 @@ namespace math {
...
@@ -27,6 +65,7 @@ namespace math {
// Then matrixA: M * K, matrixB: K * N matrixC : M * N
// Then matrixA: M * K, matrixB: K * N matrixC : M * N
// For more detailed info, please refer to
// For more detailed info, please refer to
// http://www.netlib.org/lapack/explore-html/d4/de2/sgemm_8f.html
// http://www.netlib.org/lapack/explore-html/d4/de2/sgemm_8f.html
template
<
typename
Place
,
typename
T
>
void
gemm
(
const
CBLAS_TRANSPOSE
transA
,
const
CBLAS_TRANSPOSE
transB
,
void
gemm
(
const
CBLAS_TRANSPOSE
transA
,
const
CBLAS_TRANSPOSE
transB
,
const
int
M
,
const
int
N
,
const
int
K
,
const
T
alpha
,
const
T
*
A
,
const
int
M
,
const
int
N
,
const
int
K
,
const
T
alpha
,
const
T
*
A
,
const
T
*
B
,
const
T
beta
,
T
*
C
,
platform
::
DeviceContext
*
context
);
const
T
*
B
,
const
T
beta
,
T
*
C
,
platform
::
DeviceContext
*
context
);
...
@@ -34,8 +73,8 @@ void gemm(const CBLAS_TRANSPOSE transA, const CBLAS_TRANSPOSE transB,
...
@@ -34,8 +73,8 @@ void gemm(const CBLAS_TRANSPOSE transA, const CBLAS_TRANSPOSE transB,
// matrix multiply with continuous memory
// matrix multiply with continuous memory
template
<
typename
Place
,
typename
T
>
template
<
typename
Place
,
typename
T
>
void
matmul
(
const
framework
::
Tensor
&
matrix_a
,
bool
trans_a
,
void
matmul
(
const
framework
::
Tensor
&
matrix_a
,
bool
trans_a
,
const
framework
::
Tensor
&
matrix_b
,
bool
trans_b
,
float
alpha
,
const
framework
::
Tensor
&
matrix_b
,
bool
trans_b
,
T
alpha
,
framework
::
Tensor
*
matrix_out
,
float
beta
,
framework
::
Tensor
*
matrix_out
,
T
beta
,
platform
::
DeviceContext
*
context
);
platform
::
DeviceContext
*
context
);
}
// namespace math
}
// namespace math
...
...
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