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
efa8dbc9
提交
efa8dbc9
authored
2月 09, 2017
作者:
notoraptor
浏览文件
操作
浏览文件
下载
电子邮件补丁
差异文件
Back to `GpuArray_split()`
+ some other code updates.
上级
0f4192fb
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
102 行增加
和
64 行删除
+102
-64
basic_ops.py
theano/gpuarray/basic_ops.py
+76
-45
basic.py
theano/tensor/basic.py
+26
-19
没有找到文件。
theano/gpuarray/basic_ops.py
浏览文件 @
efa8dbc9
...
@@ -1395,7 +1395,7 @@ class GpuSplit(HideC, Split):
...
@@ -1395,7 +1395,7 @@ class GpuSplit(HideC, Split):
# we reuse the perform of the CPU op, which is suitable
# we reuse the perform of the CPU op, which is suitable
def
c_code_cache_version
(
self
):
def
c_code_cache_version
(
self
):
return
(
3
,)
return
(
1
,)
def
c_headers
(
self
):
def
c_headers
(
self
):
return
[
'<numpy_compat.h>'
,
'<gpuarray_helper.h>'
]
return
[
'<numpy_compat.h>'
,
'<gpuarray_helper.h>'
]
...
@@ -1403,32 +1403,30 @@ class GpuSplit(HideC, Split):
...
@@ -1403,32 +1403,30 @@ class GpuSplit(HideC, Split):
def
c_header_dirs
(
self
):
def
c_header_dirs
(
self
):
return
[
pygpu
.
get_include
(),
os
.
path
.
dirname
(
__file__
)]
return
[
pygpu
.
get_include
(),
os
.
path
.
dirname
(
__file__
)]
def
c_support_code
(
self
):
# We want to hide the definition of this method from tensor.Split.
return
''
def
c_code
(
self
,
node
,
name
,
inputs
,
outputs
,
sub
):
def
c_code
(
self
,
node
,
name
,
inputs
,
outputs
,
sub
):
if
self
.
len_splits
==
0
:
# There are no outputs, then nothing to do.
return
''
x
,
axis
,
splits
=
inputs
x
,
axis
,
splits
=
inputs
fail
=
sub
[
'fail'
]
fail
=
sub
[
'fail'
]
x_typecode
=
pygpu
.
gpuarray
.
dtype_to_typecode
(
node
.
inputs
[
0
]
.
dtype
)
splits_dtype
=
node
.
inputs
[
2
]
.
type
.
dtype_specs
()[
1
]
splits_dtype
=
node
.
inputs
[
2
]
.
type
.
dtype_specs
()[
1
]
axis_dtype
=
node
.
inputs
[
1
]
.
type
.
dtype_specs
()[
1
]
axis_dtype
=
node
.
inputs
[
1
]
.
type
.
dtype_specs
()[
1
]
expected_splits_count
=
self
.
len_splits
expected_splits_count
=
self
.
len_splits
codes_for_checking_outputs
=
[]
# filled later.
code_for_updating_output
=
[]
# filled later.
code_for_splitting
=
[]
# filled later.
code_if_sync
=
[]
# filled later.
code_if_sync
=
[]
# filled later.
full_code_for_checking_outputs
=
''
# defined later.
full_code_for_updating_output
=
''
# defined later.
full_code_for_splitting
=
''
# defined later.
full_code_if_sync
=
''
# defined later.
full_code_if_sync
=
''
# defined later.
main_code
=
"""
main_code
=
"""
ga_order x_order = (
%(x)
s->ga.flags & GA_C_ORDER) ? GA_C_ORDER : GA_F_ORDER;
int ndim = PyGpuArray_NDIM(
%(x)
s);
int ndim = PyGpuArray_NDIM(
%(x)
s);
int axis = (int)(*(
%(axis_dtype)
s*)PyArray_GETPTR1(
%(axis)
s, 0));
int axis = (int)(*(
%(axis_dtype)
s*)PyArray_GETPTR1(
%(axis)
s, 0));
int splits_count = PyArray_SIZE(
%(splits)
s);
int splits_count = PyArray_DIM(
%(splits)
s, 0);
size_t len_along_axis, sum_of_splits = 0, current_split_start = 0;
size_t len_along_axis, sum_of_splits = 0;
%(splits_dtype)
s current_split_length = 0;
%(splits_dtype)
s current_split_length;
GpuArray view;
size_t* split_points = NULL;
GpuArray* split_views = NULL;
GpuArray** split_views_pointers = NULL;
int i;
int i;
/* Check inputs. */
/* Check inputs. */
...
@@ -1460,54 +1458,87 @@ class GpuSplit(HideC, Split):
...
@@ -1460,54 +1458,87 @@ class GpuSplit(HideC, Split):
%(fail)
s
%(fail)
s
}
}
/* Check outputs. */
/* Compute splits views. */
%(full_code_for_checking_outputs)
s
/* Compute splits. */
split_points = (size_t*) malloc((splits_count - 1) * sizeof(size_t));
if (GpuArray_view(&view, &
%(x)
s->ga) != GA_NO_ERROR) {
if (split_points == NULL) {
PyErr_SetString(PyExc_RuntimeError, "GpuSplit: unable to create a view of the input.");
PyErr_NoMemory();
%(fail)
s
}
split_points[0] = (size_t) (* (
%(splits_dtype)
s*) PyArray_GETPTR1(
%(splits)
s, 0) );
for(i = 1; i < splits_count - 1; ++i) {
split_points[i] = split_points[i - 1] + (size_t) (* (
%(splits_dtype)
s*) PyArray_GETPTR1(
%(splits)
s, i) );
}
split_views = (GpuArray*) malloc(splits_count * sizeof(GpuArray));
split_views_pointers = (GpuArray**) malloc(splits_count * sizeof(GpuArray*));
if (split_views == NULL || split_views_pointers == NULL) {
PyErr_NoMemory();
free(split_views_pointers);
free(split_views);
free(split_points);
%(fail)
s
}
for (i = 0; i < splits_count; ++i) {
split_views_pointers[i] = split_views + i;
}
if (GpuArray_split(split_views_pointers, &
%(x)
s->ga, splits_count - 1, split_points, axis) != GA_NO_ERROR) {
PyErr_SetString(PyExc_RuntimeError, "GpuSplit: unable to compute split.");
for (i = 0; i < splits_count; ++i) {
GpuArray_clear(split_views_pointers[i]);
}
free(split_views_pointers);
free(split_views);
free(split_points);
%(fail)
s
%(fail)
s
}
}
%(full_code_for_splitting)
s
/* Put split views into outputs. */
%(full_code_for_updating_output)
s
/* Free memory. */
/* Free memory. */
GpuArray_clear(&view);
for (i = 0; i < splits_count; ++i) {
GpuArray_clear(split_views_pointers[i]);
}
free(split_views_pointers);
free(split_views);
free(split_points);
/* Code added if synchronization is enabled. */
/* Code added if synchronization is enabled. */
%(full_code_if_sync)
s
%(full_code_if_sync)
s
"""
"""
for
split_index
,
output
in
enumerate
(
outputs
):
for
split_index
,
output
in
enumerate
(
outputs
):
# When checking output, we allocate a PyGpuArrayObject with 0 dims
# (that is, an object with as few memory as possible), as its GpuArray field
# will be cleared and re-used as a view during split operations.
codes_for_checking_outputs
.
append
(
"""
if (theano_prep_output(&
%(output)
s, 0, NULL,
%(x_typecode)
s, x_order,
%(x)
s->context) != 0) {
PyErr_SetString(PyExc_RuntimeError, "GpuSplit: unable to prepare an output.");
%(fail)
s
}
"""
%
locals
())
code_for_splitting
.
append
(
"""
code_for_updating_output
.
append
(
"""
current_split_length = * (
%(splits_dtype)
s*) PyArray_GETPTR1(
%(splits)
s,
%(split_index)
s);
Py_XDECREF(
%(output)
s);
view.offset = PyGpuArray_STRIDE(
%(x)
s, axis) * current_split_start;
%(output)
s = pygpu_fromgpudata(
view.dimensions[axis] = current_split_length;
split_views[
%(split_index)
s].data,
GpuArray_fix_flags(&view);
split_views[
%(split_index)
s].offset,
GpuArray_clear(&
%(output)
s->ga);
split_views[
%(split_index)
s].typecode,
if (GpuArray_view(&
%(output)
s->ga, &view) != GA_NO_ERROR) {
split_views[
%(split_index)
s].nd,
PyErr_SetString(PyExc_RuntimeError, "GpuSplit: unable to transfer a view into an output.");
split_views[
%(split_index)
s].dimensions,
GpuArray_clear(&view);
split_views[
%(split_index)
s].strides,
%(x)
s->context,
1, // output is writable
Py_None, Py_None
);
if (
%(output)
s == NULL) {
PyErr_SetString(PyExc_RuntimeError, "GpuSplit: unable to update an output from a split view.");
for (i = 0; i < splits_count; ++i) {
GpuArray_clear(split_views_pointers[i]);
}
free(split_views_pointers);
free(split_views);
free(split_points);
%(fail)
s
%(fail)
s
};
}
current_split_start += current_split_length;
"""
%
locals
())
"""
%
locals
())
if
config
.
gpuarray
.
sync
:
if
config
.
gpuarray
.
sync
:
code_if_sync
.
append
(
"GpuArray_sync(&
%(output)
s->ga);"
%
locals
())
code_if_sync
.
append
(
"GpuArray_sync(&
%(output)
s->ga);"
%
locals
())
full_code_for_checking_outputs
=
'
\r\n
'
.
join
(
codes_for_checking_outputs
)
full_code_for_updating_output
=
'
\n
'
.
join
(
code_for_updating_output
)
full_code_for_splitting
=
'
\r\n
'
.
join
(
code_for_splitting
)
full_code_if_sync
=
'
\n
'
.
join
(
code_if_sync
)
full_code_if_sync
=
'
\r\n
'
.
join
(
code_if_sync
)
return
main_code
%
locals
()
return
main_code
%
locals
()
...
...
theano/tensor/basic.py
浏览文件 @
efa8dbc9
...
@@ -3779,7 +3779,7 @@ class Split(Op):
...
@@ -3779,7 +3779,7 @@ class Split(Op):
return
self
.
make_node
(
eval_points
[
0
],
*
inputs
[
1
:])
.
outputs
return
self
.
make_node
(
eval_points
[
0
],
*
inputs
[
1
:])
.
outputs
def
c_code_cache_version
(
self
):
def
c_code_cache_version
(
self
):
return
(
2
,)
return
(
1
,)
def
c_support_code
(
self
):
def
c_support_code
(
self
):
return
"""
return
"""
...
@@ -3787,25 +3787,27 @@ class Split(Op):
...
@@ -3787,25 +3787,27 @@ class Split(Op):
int split_output_shape_is_correct (
int split_output_shape_is_correct (
PyArrayObject* output, PyArrayObject* array_to_split, int axis_to_split, npy_intp split_size
PyArrayObject* output, PyArrayObject* array_to_split, int axis_to_split, npy_intp split_size
) {
) {
if (PyArray_NDIM(output) == PyArray_NDIM(array_to_split)) {
return
int i;
PyArray_NDIM(output) == PyArray_NDIM(array_to_split)
for (i = 0; i < axis_to_split; ++i) {
&& memcmp(
if (PyArray_DIM(output, i) != PyArray_DIM(array_to_split, i)) {
PyArray_DIMS(output),
return 0;
PyArray_DIMS(array_to_split),
}
axis_to_split * sizeof(npy_intp)
}
) == 0
for (i = axis_to_split + 1; i < PyArray_NDIM(array_to_split); ++i) {
&& memcmp(
if (PyArray_DIM(output, i) != PyArray_DIM(array_to_split, i)) {
PyArray_DIMS(output) + axis_to_split + 1,
return 0;
PyArray_DIMS(array_to_split) + axis_to_split + 1,
}
(PyArray_NDIM(array_to_split) - axis_to_split - 1) * sizeof(npy_intp)
}
) == 0
return split_size == PyArray_DIM(output, axis_to_split);
&& split_size == PyArray_DIM(output, axis_to_split);
}
return 0;
}
}
"""
"""
def
c_code
(
self
,
node
,
name
,
inputs
,
outputs
,
sub
):
def
c_code
(
self
,
node
,
name
,
inputs
,
outputs
,
sub
):
if
self
.
len_splits
==
0
:
# There are no outputs, then nothing to do.
return
''
x
,
axis
,
splits
=
inputs
x
,
axis
,
splits
=
inputs
fail
=
sub
[
'fail'
]
fail
=
sub
[
'fail'
]
x_typenum
=
numpy
.
dtype
(
node
.
inputs
[
0
]
.
dtype
)
.
num
x_typenum
=
numpy
.
dtype
(
node
.
inputs
[
0
]
.
dtype
)
.
num
...
@@ -3821,7 +3823,7 @@ class Split(Op):
...
@@ -3821,7 +3823,7 @@ class Split(Op):
main_code
=
"""
main_code
=
"""
int ndim = PyArray_NDIM(
%(x)
s);
int ndim = PyArray_NDIM(
%(x)
s);
int axis = (int)(*(
%(axis_dtype)
s*)PyArray_GETPTR1(
%(axis)
s, 0));
int axis = (int)(*(
%(axis_dtype)
s*)PyArray_GETPTR1(
%(axis)
s, 0));
int splits_count = PyArray_
SIZE(
%(splits)
s
);
int splits_count = PyArray_
DIM(
%(splits)
s, 0
);
npy_intp len_along_axis, sum_of_splits = 0, current_split_length = 0, current_split_start = 0;
npy_intp len_along_axis, sum_of_splits = 0, current_split_length = 0, current_split_start = 0;
npy_intp* split_dims = NULL;
npy_intp* split_dims = NULL;
PyObject* split_view = NULL;
PyObject* split_view = NULL;
...
@@ -3862,6 +3864,11 @@ class Split(Op):
...
@@ -3862,6 +3864,11 @@ class Split(Op):
/* Check outputs. */
/* Check outputs. */
split_dims = (npy_intp*) malloc(ndim * sizeof(npy_intp));
split_dims = (npy_intp*) malloc(ndim * sizeof(npy_intp));
if (split_dims == NULL) {
PyErr_NoMemory();
%(fail)
s
}
memcpy(split_dims, PyArray_DIMS(
%(x)
s), ndim * sizeof(npy_intp));
memcpy(split_dims, PyArray_DIMS(
%(x)
s), ndim * sizeof(npy_intp));
%(full_code_for_checking_outputs)
s
%(full_code_for_checking_outputs)
s
...
@@ -3916,8 +3923,8 @@ class Split(Op):
...
@@ -3916,8 +3923,8 @@ class Split(Op):
current_split_start += current_split_length;
current_split_start += current_split_length;
"""
%
locals
())
"""
%
locals
())
full_code_for_checking_outputs
=
'
\
r\
n
'
.
join
(
codes_for_checking_outputs
)
full_code_for_checking_outputs
=
'
\n
'
.
join
(
codes_for_checking_outputs
)
full_code_for_splitting
=
'
\
r\
n
'
.
join
(
codes_for_splitting
)
full_code_for_splitting
=
'
\n
'
.
join
(
codes_for_splitting
)
return
main_code
%
locals
()
return
main_code
%
locals
()
...
...
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