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
d12a03be
提交
d12a03be
authored
9月 12, 2014
作者:
Frédéric Bastien
浏览文件
操作
浏览文件
下载
差异文件
Merge pull request #1936 from daemonmaker/issue1569
Added to test to determine the behavior of the BLAS and intialize the me...
上级
0f1c5ac3
7790b329
隐藏空白字符变更
内嵌
并排
正在显示
2 个修改的文件
包含
142 行增加
和
9 行删除
+142
-9
blas_c.py
theano/tensor/blas_c.py
+124
-7
test_blas_c.py
theano/tensor/tests/test_blas_c.py
+18
-2
没有找到文件。
theano/tensor/blas_c.py
浏览文件 @
d12a03be
import
numpy
from
theano
import
config
from
theano
import
config
from
theano.tensor.opt
import
in2out
from
theano.tensor.opt
import
in2out
...
@@ -5,6 +7,8 @@ from theano.tensor.blas import ldflags, blas_header_text, blas_header_version
...
@@ -5,6 +7,8 @@ from theano.tensor.blas import ldflags, blas_header_text, blas_header_version
from
theano.tensor.blas
import
blas_optdb
,
optdb
,
local_optimizer
,
EquilibriumOptimizer
from
theano.tensor.blas
import
blas_optdb
,
optdb
,
local_optimizer
,
EquilibriumOptimizer
from
theano.tensor.blas
import
Ger
,
ger
,
ger_destructive
from
theano.tensor.blas
import
Ger
,
ger
,
ger_destructive
from
theano.tensor.blas
import
Gemv
,
gemv_inplace
,
gemv_no_inplace
from
theano.tensor.blas
import
Gemv
,
gemv_inplace
,
gemv_no_inplace
from
theano.tensor
import
basic
as
T
import
theano.compile
class
BaseBLAS
(
object
):
class
BaseBLAS
(
object
):
...
@@ -280,13 +284,13 @@ def make_c_ger_destructive(node):
...
@@ -280,13 +284,13 @@ def make_c_ger_destructive(node):
####### ####### #######
####### ####### #######
def
gemv_c_code
(
aa
,
xx
,
yy
,
zz
,
alpha
,
beta
,
destructive
,
fail
):
def
gemv_c_code
(
aa
,
xx
,
yy
,
zz
,
alpha
,
beta
,
destructive
,
fail
,
force_init_beta
=
False
):
"""
"""
zz <- beta * aa + alpha * dot(xx, yy)
zz <- beta * aa + alpha * dot(xx, yy)
where xx is a matrix, yy and aa are vectors (ergo zz is vector)
where xx is a matrix, yy and aa are vectors (ergo zz is vector)
"""
"""
return
"""
code
=
"""
int elemsize ;
int elemsize ;
float fbeta;
float fbeta;
...
@@ -378,7 +382,7 @@ def gemv_c_code(aa, xx, yy, zz, alpha, beta, destructive, fail):
...
@@ -378,7 +382,7 @@ def gemv_c_code(aa, xx, yy, zz, alpha, beta, destructive, fail):
zoutdata[Zi*i] = fbeta * zdata[Ai*i];
zoutdata[Zi*i] = fbeta * zdata[Ai*i];
}
}
}
}
else if (PyArray_DESCR(
%(
xx
)
s)->type_num == NPY_DOUBLE)
else if (PyArray_DESCR(
%(
zz
)
s)->type_num == NPY_DOUBLE)
{
{
double * zoutdata = (double*) PyArray_DATA(
%(zz)
s);
double * zoutdata = (double*) PyArray_DATA(
%(zz)
s);
const double * zdata = (double*)PyArray_DATA(
%(aa)
s);
const double * zdata = (double*)PyArray_DATA(
%(aa)
s);
...
@@ -397,6 +401,40 @@ def gemv_c_code(aa, xx, yy, zz, alpha, beta, destructive, fail):
...
@@ -397,6 +401,40 @@ def gemv_c_code(aa, xx, yy, zz, alpha, beta, destructive, fail):
}
}
fbeta = dbeta = 1.0;
fbeta = dbeta = 1.0;
}
}
else if (
%(force_init_beta)
d)
{
if (PyArray_CHKFLAGS(
%(zz)
s, NPY_ARRAY_C_CONTIGUOUS))
{
memset((void *)PyArray_DATA(
%(zz)
s), 0, PyArray_SIZE(
%(zz)
s)*PyArray_ITEMSIZE(
%(zz)
s));
}
else
{
if (PyArray_DESCR(
%(zz)
s)->type_num == NPY_FLOAT)
{
float *zoutdata = (float *)PyArray_DATA(
%(zz)
s);
int Zi = PyArray_STRIDES(
%(zz)
s)[0]/sizeof(float);
for (int i = 0; i < PyArray_DIMS(
%(aa)
s)[0]; ++i)
{
zoutdata[Zi*i] = 0.0f;
}
}
else if (PyArray_DESCR(
%(zz)
s)->type_num == NPY_DOUBLE)
{
double *zoutdata = (double *)PyArray_DATA(
%(zz)
s);
int Zi = PyArray_STRIDES(
%(zz)
s)[0]/sizeof(double);
for (int i = 0; i < PyArray_DIMS(
%(aa)
s)[0]; ++i)
{
zoutdata[Zi*i] = 0.0;
}
}
else
{
PyErr_SetString(PyExc_AssertionError,
"neither float nor double dtype");
%(fail)
s
}
}
}
}
}
else
else
{
{
...
@@ -566,24 +604,81 @@ def gemv_c_code(aa, xx, yy, zz, alpha, beta, destructive, fail):
...
@@ -566,24 +604,81 @@ def gemv_c_code(aa, xx, yy, zz, alpha, beta, destructive, fail):
}
}
}
}
"""
%
locals
()
"""
return
code
%
locals
()
class
CGemv
(
BaseBLAS
,
Gemv
):
class
CGemv
(
BaseBLAS
,
Gemv
):
def
__init__
(
self
,
inplace
,
force_init_beta
=
False
):
super
(
CGemv
,
self
)
.
__init__
(
inplace
)
self
.
force_init_beta
=
force_init_beta
def
c_code
(
self
,
node
,
name
,
inp
,
out
,
sub
):
def
c_code
(
self
,
node
,
name
,
inp
,
out
,
sub
):
aa
,
alpha
,
xx
,
yy
,
beta
=
inp
aa
,
alpha
,
xx
,
yy
,
beta
=
inp
zz
,
=
out
zz
,
=
out
code
=
gemv_c_code
(
code
=
gemv_c_code
(
aa
,
xx
,
yy
,
zz
,
alpha
,
beta
,
aa
,
xx
,
yy
,
zz
,
alpha
,
beta
,
destructive
=
int
(
self
.
inplace
),
destructive
=
int
(
self
.
inplace
),
fail
=
sub
[
'fail'
])
fail
=
sub
[
'fail'
],
force_init_beta
=
self
.
force_init_beta
)
return
code
return
code
def
c_code_cache_version
(
self
):
def
c_code_cache_version
(
self
):
return
(
1
0
,
blas_header_version
())
return
(
1
1
,
blas_header_version
())
cgemv_inplace
=
CGemv
(
inplace
=
True
)
cgemv_inplace
=
CGemv
(
inplace
=
True
)
cgemv_no_inplace
=
CGemv
(
inplace
=
False
)
cgemv_no_inplace
=
CGemv
(
inplace
=
False
)
def
check_force_gemv_init
():
if
check_force_gemv_init
.
_force_init_beta
is
None
:
"""
Test issue 1569.
Namely when evaulating
beta*aa + alpha*dot(xx, yy)
where we set aa = betas = zeros of the correct dimensions we do not
actually set aa = zeros and instead let the BLAS perform beta*aa with
uninitialized memory for speed. Occasionally the memory contains values
that are equivalent to NaN in which case the product beta*aa contains
NaN's for correctly implemented BLAS libraries. In this situation, since
we are introducing the NaN's, we need to test whether the BLAS performs
correctly. If it *does*, i.e. it actually performs the multiplication
beta*aa which will result in NaN's in the result, then we need intialize
the memory to zeros.
"""
aa
=
T
.
vector
(
'aa'
)
yy
=
T
.
vector
(
'yy'
)
xx
=
T
.
matrix
(
'xx'
)
f
=
theano
.
function
(
[
aa
,
yy
,
xx
],
gemv_no_inplace
(
aa
,
1.
,
xx
,
yy
,
0.
),
theano
.
compile
.
Mode
(
optimizer
=
'fast_compile'
)
)
# Here we introduce NaNs into the data, if they are returned by the BLAS
# then we want gemv_c_code to initiliaze the memory to 0 so that we
# don't inadvertantly introduce NaNs to the users data.
aa_data
=
numpy
.
array
(
float
(
'NaN'
)
*
numpy
.
ones
((
2
,)),
dtype
=
theano
.
config
.
floatX
)
yy_data
=
numpy
.
array
(
numpy
.
ones
((
2
,))
*
2
,
dtype
=
theano
.
config
.
floatX
)
xx_data
=
numpy
.
array
(
numpy
.
ones
((
2
,
2
)),
dtype
=
theano
.
config
.
floatX
)
zz
=
f
(
aa_data
,
yy_data
,
xx_data
)
check_force_gemv_init
.
_force_init_beta
=
numpy
.
isnan
(
zz
)
.
any
()
return
check_force_gemv_init
.
_force_init_beta
check_force_gemv_init
.
_force_init_beta
=
None
@local_optimizer
([
gemv_inplace
,
gemv_no_inplace
])
@local_optimizer
([
gemv_inplace
,
gemv_no_inplace
])
def
use_c_gemv
(
node
):
def
use_c_gemv
(
node
):
...
@@ -592,7 +687,29 @@ def use_c_gemv(node):
...
@@ -592,7 +687,29 @@ def use_c_gemv(node):
# Only float32 and float64 are supported for now.
# Only float32 and float64 are supported for now.
if
(
node
.
op
==
gemv_no_inplace
and
if
(
node
.
op
==
gemv_no_inplace
and
node
.
outputs
[
0
]
.
dtype
in
[
'float32'
,
'float64'
]):
node
.
outputs
[
0
]
.
dtype
in
[
'float32'
,
'float64'
]):
return
[
CGemv
(
inplace
=
False
)(
*
node
.
inputs
)]
"""
We want to maintain the behavoir of any operation that the user adds
even if it results in NaNs. However we do not want optimizations to
introduce NaNs.
GEMV is not always implemented consistenly across BLAS libraries.
Sometimes, when beta is 0, they do not perform the multiplication with
beta. Other implmentations do. This can cause problems for the inplace
GEMV implementation if NaNs happen to be in the newly allocated but
uninitalized memory. When the multiplication is not done we do not need
to initialize the output memory resulting in a speed up. Otherwise we
must initialize the memory to avoid introducing NaN's in the output
that weren't in the original graph.
The following check determines whether the output memory needs to be
initiliazed. It is done here, as opposed to in global scope, because
the setup has not been completed at that time and therefore the check
cannot be performed at that time.
"""
force_init_beta
=
check_force_gemv_init
()
return
[
CGemv
(
inplace
=
False
,
force_init_beta
=
force_init_beta
)(
*
node
.
inputs
)]
if
(
node
.
op
==
gemv_inplace
and
if
(
node
.
op
==
gemv_inplace
and
node
.
outputs
[
0
]
.
dtype
in
[
'float32'
,
'float64'
]):
node
.
outputs
[
0
]
.
dtype
in
[
'float32'
,
'float64'
]):
return
[
CGemv
(
inplace
=
True
)(
*
node
.
inputs
)]
return
[
CGemv
(
inplace
=
True
)(
*
node
.
inputs
)]
...
...
theano/tensor/tests/test_blas_c.py
浏览文件 @
d12a03be
...
@@ -15,6 +15,8 @@ from theano.tensor.blas_c import CGemv
...
@@ -15,6 +15,8 @@ from theano.tensor.blas_c import CGemv
from
theano.tensor.blas_scipy
import
ScipyGer
from
theano.tensor.blas_scipy
import
ScipyGer
from
theano.tensor.blas
import
Gemv
from
theano.tensor.blas
import
Gemv
from
theano.tensor.blas_c
import
check_force_gemv_init
from
theano.tests
import
unittest_tools
from
theano.tests
import
unittest_tools
from
theano.tests.unittest_tools
import
TestOptimizationMixin
from
theano.tests.unittest_tools
import
TestOptimizationMixin
...
@@ -137,7 +139,10 @@ class TestCGemv(TestCase, TestOptimizationMixin):
...
@@ -137,7 +139,10 @@ class TestCGemv(TestCase, TestOptimizationMixin):
# Assert that the dot was optimized somehow
# Assert that the dot was optimized somehow
self
.
assertFunctionContains0
(
f
,
tensor
.
dot
)
self
.
assertFunctionContains0
(
f
,
tensor
.
dot
)
self
.
assertFunctionContains1
(
f
,
CGemv
(
True
))
self
.
assertFunctionContains1
(
f
,
CGemv
(
inplace
=
True
,
force_init_beta
=
True
)
)
# Assert they produce the same output
# Assert they produce the same output
assert
numpy
.
allclose
(
f
(
self
.
xval
,
self
.
Aval
),
assert
numpy
.
allclose
(
f
(
self
.
xval
,
self
.
Aval
),
...
@@ -155,7 +160,10 @@ class TestCGemv(TestCase, TestOptimizationMixin):
...
@@ -155,7 +160,10 @@ class TestCGemv(TestCase, TestOptimizationMixin):
# Assert that the dot was optimized somehow
# Assert that the dot was optimized somehow
self
.
assertFunctionContains0
(
f
,
tensor
.
dot
)
self
.
assertFunctionContains0
(
f
,
tensor
.
dot
)
self
.
assertFunctionContains1
(
f
,
CGemv
(
True
))
self
.
assertFunctionContains1
(
f
,
CGemv
(
inplace
=
True
,
force_init_beta
=
True
)
)
# Assert they produce the same output
# Assert they produce the same output
assert
numpy
.
allclose
(
f
(
self
.
Aval
,
self
.
yval
),
assert
numpy
.
allclose
(
f
(
self
.
Aval
,
self
.
yval
),
...
@@ -164,6 +172,14 @@ class TestCGemv(TestCase, TestOptimizationMixin):
...
@@ -164,6 +172,14 @@ class TestCGemv(TestCase, TestOptimizationMixin):
assert
numpy
.
allclose
(
f
(
self
.
Aval
[::
-
1
,
::
-
1
],
self
.
yval
),
assert
numpy
.
allclose
(
f
(
self
.
Aval
[::
-
1
,
::
-
1
],
self
.
yval
),
numpy
.
dot
(
self
.
Aval
[::
-
1
,
::
-
1
],
self
.
yval
))
numpy
.
dot
(
self
.
Aval
[::
-
1
,
::
-
1
],
self
.
yval
))
def
test_force_gemv_init
(
self
):
if
check_force_gemv_init
():
sys
.
stderr
.
write
(
"WARNING: The current BLAS requires Theano to initialize"
+
" memory for some GEMV calls which will result in a minor"
+
" degradation in performance for such calls."
)
def
t_gemv1
(
self
,
m_shp
):
def
t_gemv1
(
self
,
m_shp
):
''' test vector2 + dot(matrix, vector1) '''
''' test vector2 + dot(matrix, vector1) '''
rng
=
numpy
.
random
.
RandomState
(
unittest_tools
.
fetch_seed
())
rng
=
numpy
.
random
.
RandomState
(
unittest_tools
.
fetch_seed
())
...
...
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