Skip to content
项目
群组
代码片段
帮助
当前项目
正在载入...
登录 / 注册
切换导航面板
P
pytensor
项目
项目
详情
活动
周期分析
仓库
仓库
文件
提交
分支
标签
贡献者
图表
比较
统计图
议题
0
议题
0
列表
看板
标记
里程碑
合并请求
0
合并请求
0
CI / CD
CI / CD
流水线
作业
日程
统计图
Wiki
Wiki
代码片段
代码片段
成员
成员
折叠边栏
关闭边栏
活动
图像
聊天
创建新问题
作业
提交
问题看板
Open sidebar
testgroup
pytensor
Commits
f57ea7bd
提交
f57ea7bd
authored
6月 30, 2010
作者:
James Bergstra
浏览文件
操作
浏览文件
下载
电子邮件补丁
差异文件
modified gen_graph for benchmarking elementwise stuff to make stdout comprehensible
上级
00587f67
显示空白字符变更
内嵌
并排
正在显示
1 个修改的文件
包含
63 行增加
和
38 行删除
+63
-38
gen_graph.py
benchmark/numexpr/gen_graph.py
+63
-38
没有找到文件。
benchmark/numexpr/gen_graph.py
浏览文件 @
f57ea7bd
...
@@ -3,8 +3,11 @@ import numexpr as ne
...
@@ -3,8 +3,11 @@ import numexpr as ne
import
timeit
import
timeit
import
theano
import
theano
import
theano.tensor
as
T
import
theano.tensor
as
T
import
pylab
try
:
import
matplotlib.pyplot
as
pyplot
import
pylab
import
matplotlib.pyplot
as
pyplot
except
ImportError
:
pass
def
timeit_2vector_theano
(
init
,
nb_element
=
1e6
,
nb_repeat
=
3
,
nb_call
=
int
(
1e2
),
expr
=
"a**2 + b**2 + 2*a*b"
):
def
timeit_2vector_theano
(
init
,
nb_element
=
1e6
,
nb_repeat
=
3
,
nb_call
=
int
(
1e2
),
expr
=
"a**2 + b**2 + 2*a*b"
):
t3
=
timeit
.
Timer
(
"tf(av,bv)"
,
t3
=
timeit
.
Timer
(
"tf(av,bv)"
,
...
@@ -23,7 +26,13 @@ tf= theano.function([a,b],%(expr)s)
...
@@ -23,7 +26,13 @@ tf= theano.function([a,b],%(expr)s)
)
)
ret
=
t3
.
repeat
(
nb_repeat
,
nb_call
)
ret
=
t3
.
repeat
(
nb_repeat
,
nb_call
)
return
np
.
asarray
(
ret
)
return
np
.
asarray
(
ret
)
def
timeit_2vector
(
nb_element
=
1e6
,
nb_repeat
=
3
,
nb_call
=
int
(
1e2
),
expr
=
"a**2 + b**2 + 2*a*b"
,
do_unalign
=
False
,
do_amd
=
True
):
def
timeit_2vector
(
nb_element
=
1e6
,
nb_repeat
=
3
,
nb_call
=
int
(
1e2
),
expr
=
"a**2 + b**2 + 2*a*b"
,
do_unalign
=
False
,
do_amd
=
True
):
"""Returns a dictionary whose keys are implementations ('numpy', 'numexpr', 'theano', etc.)
and whose values are numpy arrays of times taken to evalute the given problem.
"""
rval
=
dict
()
print
print
print
"timeit_2vector(nb_element=
%(nb_element)
s,nb_repeat=
%(nb_repeat)
s,nb_call=
%(nb_call)
s, expr=
%(expr)
s, do_unalign=
%(do_unalign)
s)"
%
locals
()
print
"timeit_2vector(nb_element=
%(nb_element)
s,nb_repeat=
%(nb_repeat)
s,nb_call=
%(nb_call)
s, expr=
%(expr)
s, do_unalign=
%(do_unalign)
s)"
%
locals
()
...
@@ -32,60 +41,74 @@ def timeit_2vector(nb_element=1e6, nb_repeat=3, nb_call=int(1e2), expr="a**2 + b
...
@@ -32,60 +41,74 @@ def timeit_2vector(nb_element=1e6, nb_repeat=3, nb_call=int(1e2), expr="a**2 + b
else
:
else
:
init
=
"import numpy as np; a = np.arange(
%(nb_element)
s);b = np.arange(
%(nb_element)
s)"
%
locals
()
init
=
"import numpy as np; a = np.arange(
%(nb_element)
s);b = np.arange(
%(nb_element)
s)"
%
locals
()
t1
=
timeit
.
Timer
(
"
%(expr)
s"
%
locals
(),
"from numpy import exp;
%(init)
s"
%
locals
())
t1
=
timeit
.
Timer
(
"
%(expr)
s"
%
locals
(),
"from numpy import exp;
%(init)
s"
%
locals
())
ret1
=
t1
.
repeat
(
nb_repeat
,
nb_call
)
numpy_times
=
np
.
asarray
(
t1
.
repeat
(
nb_repeat
,
nb_call
))
ret1
=
np
.
asarray
(
ret1
)
print
"NumPy time: each time="
,
numpy_times
,
"min_time="
,
numpy_times
.
min
()
print
"NumPy time"
,
ret1
,
ret1
.
min
()
rval
[
'numpy'
]
=
numpy_times
t2
=
timeit
.
Timer
(
"""ne.evaluate("
%(expr)
s")"""
%
locals
(),
t2
=
timeit
.
Timer
(
"""ne.evaluate("
%(expr)
s")"""
%
locals
(),
"import numexpr as ne;
%(init)
s"
%
locals
())
"import numexpr as ne;
%(init)
s"
%
locals
())
ret2
=
t2
.
repeat
(
nb_repeat
,
nb_call
)
numexpr_times
=
np
.
asarray
(
t2
.
repeat
(
nb_repeat
,
nb_call
)
)
r
et2
=
np
.
asarray
(
ret2
)
r
val
[
'numexpr'
]
=
numexpr_times
print
"Numexpr time
"
,
ret2
,
ret2
.
min
()
print
"Numexpr time
: each time="
,
numexpr_times
,
'min_time='
,
numexpr_times
.
min
()
theano
.
config
.
lib
.
amdlibm
=
False
theano
.
config
.
lib
.
amdlibm
=
False
ret3
=
timeit_2vector_theano
(
init
,
nb_element
,
nb_repeat
,
nb_call
,
expr
)
theano_times
=
timeit_2vector_theano
(
init
,
nb_element
,
nb_repeat
,
nb_call
,
expr
)
print
"Theano time"
,
ret3
,
ret3
.
min
()
print
"Theano time: each time="
,
theano_times
,
'min_time='
,
theano_times
.
min
()
rval
[
'theano'
]
=
theano_times
if
do_amd
:
if
do_amd
:
theano
.
config
.
lib
.
amdlibm
=
True
theano
.
config
.
lib
.
amdlibm
=
True
ret4
=
timeit_2vector_theano
(
init
,
nb_element
,
nb_repeat
,
nb_call
,
expr
)
theanoamd_times
=
timeit_2vector_theano
(
init
,
nb_element
,
nb_repeat
,
nb_call
,
expr
)
print
"Theano time(with amdlibm)"
,
ret3
,
ret3
.
min
()
print
"Theano+amdlibm time"
,
theanoamd_times
,
theanoamd_times
.
min
()
rval
[
'theano_amd'
]
=
theanoamd_times
print
"
Numexpr vs NumPy"
,
ret1
.
min
()
/
ret2
.
min
()
print
"
time(NumPy) / time(numexpr) = "
,
numpy_times
.
min
()
/
numexpr_times
.
min
()
print
"
Theano vs NumPy"
,
ret1
.
min
()
/
ret3
.
min
()
print
"
time(NumPy) / time(Theano)"
,
numpy_times
.
min
()
/
theano_times
.
min
()
print
"
Theano vs Numexpr"
,
ret2
.
min
()
/
ret3
.
min
()
print
"
time(numexpr) / time(Theano)"
,
numexpr_times
.
min
()
/
theano_times
.
min
()
if
do_amd
:
if
do_amd
:
print
"Theano(amdlibm) vs NumPy"
,
ret1
.
min
()
/
ret4
.
min
()
print
"time(NumPy) / time(Theano+amdlibm)"
,
numpy_times
.
min
()
/
theanoamd_times
.
min
()
print
"Theano(amdlibm) vs Numexpr"
,
ret2
.
min
()
/
ret4
.
min
()
print
"time(numexpr) / time(Theano+amdlibm)"
,
numexpr_times
.
min
()
/
theanoamd_times
.
min
()
return
ret1
,
ret2
,
ret3
,
ret4
return
rval
return
ret1
,
ret2
,
ret3
def
exec_timeit_2vector
(
expr
,
nb_call_scal
=
1
,
fname
=
None
,
do_unalign
=
False
,
do_amd
=
True
):
def
exec_timeit_2vector
(
expr
,
nb_call_scal
=
1
,
fname
=
None
,
do_unalign
=
False
,
do_amd
=
True
):
time
=
[]
#exp = [(1,100000),(1e1,100000),(1e2,100000),(1e3,100000), (5e3,50000),
#exp = [(1,100000),(1e1,100000),(1e2,100000),(1e3,100000), (5e3,50000),
exp
=
[(
1e3
,
100000
),(
5e3
,
50000
),
\
exp
=
[(
1e3
,
100000
),(
5e3
,
50000
),
\
(
1e4
,
10000
),(
5e4
,
5000
),(
1e5
,
2000
),(
1e6
,
200
),(
1e7
,
10
)
(
1e4
,
10000
),(
5e4
,
5000
),(
1e5
,
2000
),(
1e6
,
200
),(
1e7
,
10
)
]
]
exp
=
[(
1e3
,
100000
),(
5e3
,
50000
)]
runtimes
=
[]
for
nb_e
,
nb_c
in
exp
:
for
nb_e
,
nb_c
in
exp
:
time
.
append
(
timeit_2vector
(
nb_element
=
nb_e
,
nb_repeat
=
3
,
nb_call
=
nb_c
*
nb_call_scal
,
expr
=
expr
,
do_amd
=
do_amd
))
runtimes
.
append
(
timeit_2vector
(
nb_element
=
nb_e
,
nb_repeat
=
3
,
nb_call
=
nb_c
*
nb_call_scal
,
expr
=
expr
,
do_amd
=
do_amd
))
if
do_unalign
:
if
do_unalign
:
time
_unalign
=
[]
runtimes
_unalign
=
[]
for
nb_e
,
nb_c
in
exp
:
for
nb_e
,
nb_c
in
exp
:
time_unalign
.
append
(
timeit_2vector
(
nb_element
=
nb_e
,
nb_repeat
=
3
,
nb_call
=
nb_c
*
nb_call_scal
,
expr
=
expr
,
do_unalign
=
True
,
do_amd
=
do_amd
))
runtimes_unalign
.
append
(
timeit_2vector
(
nb_element
=
nb_e
,
nb_repeat
=
3
,
nb_call
=
nb_c
*
nb_call_scal
,
expr
=
expr
,
do_unalign
=
True
,
do_amd
=
do_amd
))
print
time
print
'Runtimes list = '
,
runtimes
num_speedup
=
np
.
asarray
([
t
[
0
]
.
min
()
/
t
[
1
]
.
min
()
for
t
in
time
],
"float32"
)
numexpr_speedup
=
np
.
asarray
([
t
[
'numpy'
]
.
min
()
/
t
[
'numexpr'
]
.
min
()
for
t
in
runtimes
],
"float32"
)
print
"Numexpr vs NumPy"
,
num_speedup
,
num_speedup
.
min
(),
num_speedup
.
max
()
print
"time(NumPy) / time(numexpr)"
,
theano_speedup
=
np
.
asarray
([
t
[
0
]
.
min
()
/
t
[
2
]
.
min
()
for
t
in
time
],
"float32"
)
print
numexpr_speedup
,
numexpr_speedup
.
min
(),
numexpr_speedup
.
max
()
print
"Theano vs NumPy"
,
theano_speedup
,
theano_speedup
.
min
(),
theano_speedup
.
max
()
theano_num_speedup
=
np
.
asarray
([
t
[
1
]
.
min
()
/
t
[
2
]
.
min
()
for
t
in
time
],
"float32"
)
theano_speedup
=
np
.
asarray
([
t
[
'numpy'
]
.
min
()
/
t
[
'theano'
]
.
min
()
for
t
in
runtimes
],
"float32"
)
print
"Theano vs Numexpr"
,
theano_num_speedup
,
theano_num_speedup
.
min
(),
theano_num_speedup
.
max
()
print
"time(NumPy) / time(Theano)"
,
print
theano_speedup
,
theano_speedup
.
min
(),
theano_speedup
.
max
()
theano_numexpr_speedup
=
np
.
asarray
([
t
[
'numexpr'
]
.
min
()
/
t
[
'theano'
]
.
min
()
for
t
in
runtimes
],
"float32"
)
print
"time(numexpr) / time(Theano)"
,
print
theano_numexpr_speedup
,
theano_numexpr_speedup
.
min
(),
theano_numexpr_speedup
.
max
()
if
do_amd
:
if
do_amd
:
theano_speedup2
=
np
.
asarray
([
t
[
0
]
.
min
()
/
t
[
3
]
.
min
()
for
t
in
time
],
"float32"
)
theano_speedup2
=
np
.
asarray
([
t
[
'numpy'
]
.
min
()
/
t
[
'theano_amd'
]
.
min
()
for
t
in
runtimes
],
"float32"
)
print
"Theano vs NumPy"
,
theano_speedup
,
theano_speedup
.
min
(),
theano_speedup
.
max
()
print
"time(NumPy) / time(theano+amdlibm)"
,
theano_num_speedup2
=
np
.
asarray
([
t
[
1
]
.
min
()
/
t
[
3
]
.
min
()
for
t
in
time
],
"float32"
)
print
theano_speedup
,
theano_speedup
.
min
(),
theano_speedup
.
max
()
print
"Theano vs Numexpr"
,
theano_num_speedup
,
theano_num_speedup
.
min
(),
theano_num_speedup
.
max
()
theano_numexpr_speedup2
=
np
.
asarray
([
t
[
'numexpr'
]
.
min
()
/
t
[
'theano_amd'
]
.
min
()
for
t
in
runtimes
],
"float32"
)
print
"time(numexpr) / time(theano+amdlibm)"
,
print
theano_numexpr_speedup
,
theano_numexpr_speedup
.
min
(),
theano_numexpr_speedup
.
max
()
if
'pylab'
not
in
globals
():
return
nb_calls
=
[
e
[
0
]
for
e
in
exp
]
nb_calls
=
[
e
[
0
]
for
e
in
exp
]
for
cmp
in
range
(
1
,
len
(
time
[
0
])):
for
cmp
in
range
(
1
,
len
(
time
[
0
])):
...
@@ -128,8 +151,8 @@ def execs_timeit_2vector(exprs, fname=None):
...
@@ -128,8 +151,8 @@ def execs_timeit_2vector(exprs, fname=None):
exp
=
[(
1e3
,
100000
),(
5e3
,
50000
),
\
exp
=
[(
1e3
,
100000
),(
5e3
,
50000
),
\
(
1e4
,
10000
),(
5e4
,
5000
),(
1e5
,
2000
),(
1e6
,
200
),(
1e7
,
10
)
(
1e4
,
10000
),(
5e4
,
5000
),(
1e5
,
2000
),(
1e6
,
200
),(
1e7
,
10
)
]
]
#
TO TEST UNCOMMENT THIS LINE
###
TO TEST UNCOMMENT THIS LINE
#exp = [(1,1000),(1e1,1000),(1e2,1000),]
#
exp = [(1,1000),(1e1,1000),(1e2,1000),]
times
=
[]
times
=
[]
str_expr
=
[]
str_expr
=
[]
for
g_exprs
in
exprs
:
for
g_exprs
in
exprs
:
...
@@ -143,6 +166,8 @@ def execs_timeit_2vector(exprs, fname=None):
...
@@ -143,6 +166,8 @@ def execs_timeit_2vector(exprs, fname=None):
for
nb_e
,
nb_c
in
exp
:
for
nb_e
,
nb_c
in
exp
:
time
.
append
(
timeit_2vector
(
nb_element
=
nb_e
,
nb_repeat
=
3
,
nb_call
=
nb_c
*
nb_call_scal
,
expr
=
expr
,
do_amd
=
False
))
time
.
append
(
timeit_2vector
(
nb_element
=
nb_e
,
nb_repeat
=
3
,
nb_call
=
nb_c
*
nb_call_scal
,
expr
=
expr
,
do_amd
=
False
))
times
.
append
(
time
)
times
.
append
(
time
)
if
'pylab'
not
in
globals
():
return
nb_calls
=
[
e
[
0
]
for
e
in
exp
]
nb_calls
=
[
e
[
0
]
for
e
in
exp
]
legends
=
[]
legends
=
[]
...
...
编写
预览
Markdown
格式
0%
重试
或
添加新文件
添加附件
取消
您添加了
0
人
到此讨论。请谨慎行事。
请先完成此评论的编辑!
取消
请
注册
或者
登录
后发表评论