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
Crayon鑫
Paddle
提交
3149b040
P
Paddle
项目概览
Crayon鑫
/
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
提交
3149b040
编写于
11月 08, 2021
作者:
L
lilong12
提交者:
GitHub
11月 08, 2021
浏览文件
操作
浏览文件
下载
电子邮件补丁
差异文件
Revert "Add reference to global_gather and global_scatter operators (#36985)"
This reverts commit
aef291f4
.
上级
ac1d3571
变更
2
隐藏空白更改
内联
并排
Showing
2 changed file
with
4 addition
and
6 deletion
+4
-6
paddle/fluid/operators/collective/global_gather_op.cu.cc
paddle/fluid/operators/collective/global_gather_op.cu.cc
+2
-3
paddle/fluid/operators/collective/global_scatter_op.cu.cc
paddle/fluid/operators/collective/global_scatter_op.cu.cc
+2
-3
未找到文件。
paddle/fluid/operators/collective/global_gather_op.cu.cc
浏览文件 @
3149b040
...
@@ -10,7 +10,7 @@ Unless required by applicable law or agreed to in writing, software
...
@@ -10,7 +10,7 @@ Unless required by applicable law or agreed to in writing, software
distributed under the License is distributed on an "AS IS" BASIS,
distributed under the License is distributed on an "AS IS" BASIS,
WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
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.
*/
#include "paddle/fluid/operators/collective/global_gather_op.h"
#include "paddle/fluid/operators/collective/global_gather_op.h"
...
@@ -101,8 +101,7 @@ class GlobalGatherOpCUDAKernel : public framework::OpKernel<T> {
...
@@ -101,8 +101,7 @@ class GlobalGatherOpCUDAKernel : public framework::OpKernel<T> {
auto
send_ptr
=
0
;
auto
send_ptr
=
0
;
auto
send_buf
=
x
->
data
<
T
>
();
auto
send_buf
=
x
->
data
<
T
>
();
auto
recv_buf
=
out
->
mutable_data
<
T
>
(
out_dims
,
place
);
auto
recv_buf
=
out
->
mutable_data
<
T
>
(
out_dims
,
place
);
// Taken and modified for PaddlePaddle from:
// https://github.com/laekov/fastmoe
for
(
auto
i
=
0
;
i
<
n_expert
;
++
i
)
{
for
(
auto
i
=
0
;
i
<
n_expert
;
++
i
)
{
PADDLE_ENFORCE_CUDA_SUCCESS
(
platform
::
dynload
::
ncclGroupStart
());
PADDLE_ENFORCE_CUDA_SUCCESS
(
platform
::
dynload
::
ncclGroupStart
());
for
(
auto
j
=
0
;
j
<
nranks
;
++
j
)
{
for
(
auto
j
=
0
;
j
<
nranks
;
++
j
)
{
...
...
paddle/fluid/operators/collective/global_scatter_op.cu.cc
浏览文件 @
3149b040
...
@@ -10,7 +10,7 @@ Unless required by applicable law or agreed to in writing, software
...
@@ -10,7 +10,7 @@ Unless required by applicable law or agreed to in writing, software
distributed under the License is distributed on an "AS IS" BASIS,
distributed under the License is distributed on an "AS IS" BASIS,
WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
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.
*/
#include "paddle/fluid/operators/collective/global_scatter_op.h"
#include "paddle/fluid/operators/collective/global_scatter_op.h"
...
@@ -100,8 +100,7 @@ class GlobalScatterOpCUDAKernel : public framework::OpKernel<T> {
...
@@ -100,8 +100,7 @@ class GlobalScatterOpCUDAKernel : public framework::OpKernel<T> {
auto
recv_ptr
=
0
;
auto
recv_ptr
=
0
;
auto
send_buf
=
x
->
data
<
T
>
();
auto
send_buf
=
x
->
data
<
T
>
();
auto
recv_buf
=
out
->
mutable_data
<
T
>
(
out_dims
,
place
);
auto
recv_buf
=
out
->
mutable_data
<
T
>
(
out_dims
,
place
);
// Taken and modified for PaddlePaddle from:
// https://github.com/laekov/fastmoe
for
(
auto
i
=
0
;
i
<
n_expert
;
++
i
)
{
for
(
auto
i
=
0
;
i
<
n_expert
;
++
i
)
{
PADDLE_ENFORCE_CUDA_SUCCESS
(
platform
::
dynload
::
ncclGroupStart
());
PADDLE_ENFORCE_CUDA_SUCCESS
(
platform
::
dynload
::
ncclGroupStart
());
for
(
auto
j
=
0
;
j
<
nranks
;
++
j
)
{
for
(
auto
j
=
0
;
j
<
nranks
;
++
j
)
{
...
...
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