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
机器未来
Paddle
提交
d00e1ed5
P
Paddle
项目概览
机器未来
/
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
未验证
提交
d00e1ed5
编写于
12月 29, 2017
作者:
T
Tao Luo
提交者:
GitHub
12月 29, 2017
浏览文件
操作
浏览文件
下载
差异文件
Merge pull request #7104 from tensor-tang/openblas
auto set openblas env
上级
af602673
d630d392
变更
3
隐藏空白更改
内联
并排
Showing
3 changed file
with
25 addition
and
16 deletion
+25
-16
benchmark/paddle/image/run_openblas_infer.sh
benchmark/paddle/image/run_openblas_infer.sh
+10
-6
benchmark/paddle/image/run_openblas_train.sh
benchmark/paddle/image/run_openblas_train.sh
+1
-1
paddle/scripts/submit_local.sh.in
paddle/scripts/submit_local.sh.in
+14
-9
未找到文件。
benchmark/paddle/image/run_openblas_infer.sh
浏览文件 @
d00e1ed5
...
...
@@ -8,15 +8,19 @@ function clock_to_seconds() {
}
function
infer
()
{
unset
OMP_NUM_THREADS MKL_NUM_THREADS OMP_DYNAMIC KMP_AFFINITY
topology
=
$1
layer_num
=
$2
bs
=
$3
t
hread
=
`
nproc
`
if
[
$t
hread
-gt
$bs
]
;
then
t
hread
=
$bs
t
rainers
=
`
nproc
`
if
[
$t
rainers
-gt
$bs
]
;
then
t
rainers
=
$bs
fi
log
=
"logs/infer-
${
topology
}
-
${
layer_num
}
-
${
thread
}
openblas-
${
bs
}
.log"
log
=
"logs/infer-
${
topology
}
-
${
layer_num
}
-
${
trainers
}
openblas-
${
bs
}
.log"
threads
=
$((
`
nproc
`
/
trainers
))
if
[
$threads
-eq
0
]
;
then
threads
=
1
fi
export
OPENBLAS_NUM_THREADS
=
$threads
models_in
=
"models/
${
topology
}
-
${
layer_num
}
/pass-00000/"
if
[
!
-d
$models_in
]
;
then
...
...
@@ -28,7 +32,7 @@ function infer() {
--config
=
"
${
topology
}
.py"
\
--use_mkldnn
=
False
\
--use_gpu
=
False
\
--trainer_count
=
$t
hread
\
--trainer_count
=
$t
rainers
\
--log_period
=
$log_period
\
--config_args
=
"batch_size=
${
bs
}
,layer_num=
${
layer_num
}
,is_infer=True,num_samples=256"
\
--init_model_path
=
$models_in
\
...
...
benchmark/paddle/image/run_openblas_train.sh
浏览文件 @
d00e1ed5
set
-e
function
train
()
{
unset
OMP_NUM_THREADS MKL_NUM_THREADS OMP_DYNAMIC KMP_AFFINITY
export
OPENBLAS_NUM_THREADS
=
1
topology
=
$1
layer_num
=
$2
bs
=
$3
...
...
paddle/scripts/submit_local.sh.in
浏览文件 @
d00e1ed5
...
...
@@ -71,9 +71,7 @@ function threads_config() {
# auto set OMP_NUM_THREADS and MKL_NUM_THREADS
# according to trainer_count and total processors
# only when MKL enabled
if
[
"@WITH_MKL@"
==
"OFF"
]
;
then
return
0
fi
# auto set OPENBLAS_NUM_THREADS when do not use MKL
processors
=
`
grep
"processor"
/proc/cpuinfo|sort
-u
|wc
-l
`
trainers
=
`
grep
-Eo
'trainer_count.[0-9]+'
<<<
"
$@
"
|grep
-Eo
'[0-9]+'
|xargs
`
if
[
-z
$trainers
]
;
then
...
...
@@ -83,12 +81,19 @@ function threads_config() {
if
[
$threads
-eq
0
]
;
then
threads
=
1
fi
if
[
-z
"
$OMP_NUM_THREADS
"
]
;
then
export
OMP_NUM_THREADS
=
$threads
fi
if
[
-z
"
$MKL_NUM_THREADS
"
]
;
then
export
MKL_NUM_THREADS
=
$threads
if
[
"@WITH_MKL@"
==
"ON"
]
;
then
if
[
-z
"
$OMP_NUM_THREADS
"
]
;
then
export
OMP_NUM_THREADS
=
$threads
fi
if
[
-z
"
$MKL_NUM_THREADS
"
]
;
then
export
MKL_NUM_THREADS
=
$threads
fi
else
if
[
-z
"
$OPENBLAS_NUM_THREADS
"
]
;
then
export
OPENBLAS_NUM_THREADS
=
$threads
fi
fi
}
PADDLE_CONF_HOME
=
"
$HOME
/.config/paddle"
...
...
@@ -150,7 +155,7 @@ fi
case
"
$1
"
in
"train"
)
threads_config
$@
# echo $OMP_NUM_THREADS $MKL_NUM_THREADS
# echo $OMP_NUM_THREADS $MKL_NUM_THREADS
$OPENBLAS_NUM_THREADS
${
DEBUGGER
}
$PADDLE_BIN_PATH
/paddle_trainer
${
@
:2
}
;;
"merge_model"
)
...
...
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