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
d912ebb9
提交
d912ebb9
authored
2月 21, 2012
作者:
Frederic
浏览文件
操作
浏览文件
下载
电子邮件补丁
差异文件
Put the nvcc.flags in the hard part of the Theano keys.
上级
f1ea1df3
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
6 行增加
和
3 行删除
+6
-3
NEWS.txt
NEWS.txt
+4
-0
nvcc_compiler.py
theano/sandbox/cuda/nvcc_compiler.py
+2
-3
没有找到文件。
NEWS.txt
浏览文件 @
d912ebb9
...
@@ -11,6 +11,10 @@ Since 0.5rc2
...
@@ -11,6 +11,10 @@ Since 0.5rc2
* Fix a bug with Gemv and Ger on CPU, when used on vectors with negative
* Fix a bug with Gemv and Ger on CPU, when used on vectors with negative
strides. Data was read from incorrect (and possibly uninitialized)
strides. Data was read from incorrect (and possibly uninitialized)
memory space. This bug was probably introduced in 0.5rc1.
memory space. This bug was probably introduced in 0.5rc1.
* The Theano flag "nvcc.flags" are now included in the hard part of the key.
This mean that now we recompile all modules for each value of "nvcc.flags".
This do use change the default, but if you used this flags, it was ignored
for module already compiled.
=============
=============
Release Notes
Release Notes
...
...
theano/sandbox/cuda/nvcc_compiler.py
浏览文件 @
d912ebb9
...
@@ -80,8 +80,8 @@ class NVCC_compiler():
...
@@ -80,8 +80,8 @@ class NVCC_compiler():
This args will be received by compile_str() in the preargs paramter.
This args will be received by compile_str() in the preargs paramter.
They will also be included in the "hard" part of the key module.
They will also be included in the "hard" part of the key module.
"""
"""
return
[
]
flags
=
[
flag
for
flag
in
config
.
nvcc
.
flags
.
split
(
' '
)
if
flag
]
# flags = [flag for flag in config.nvcc.flags.split(' ') if flag]
return
flags
# cuda_ndarray_cuh_hash = hash_from_file(
# cuda_ndarray_cuh_hash = hash_from_file(
# os.path.join(os.path.split(__file__)[0], 'cuda_ndarray.cuh'))
# os.path.join(os.path.split(__file__)[0], 'cuda_ndarray.cuh'))
# cuda_macro = '-DCUDA_NDARRAY_CUH=' + cuda_ndarray_cuh_hash
# cuda_macro = '-DCUDA_NDARRAY_CUH=' + cuda_ndarray_cuh_hash
...
@@ -205,7 +205,6 @@ class NVCC_compiler():
...
@@ -205,7 +205,6 @@ class NVCC_compiler():
# the -rpath option is not understood by the Microsoft linker
# the -rpath option is not understood by the Microsoft linker
for
rpath
in
rpaths
:
for
rpath
in
rpaths
:
cmd
.
extend
([
'-Xlinker'
,
','
.
join
([
'-rpath'
,
rpath
])])
cmd
.
extend
([
'-Xlinker'
,
','
.
join
([
'-rpath'
,
rpath
])])
cmd
.
extend
([
flag
for
flag
in
config
.
nvcc
.
flags
.
split
(
' '
)
if
flag
])
cmd
.
extend
(
'-I
%
s'
%
idir
for
idir
in
include_dirs
)
cmd
.
extend
(
'-I
%
s'
%
idir
for
idir
in
include_dirs
)
cmd
.
extend
([
'-o'
,
lib_filename
])
cmd
.
extend
([
'-o'
,
lib_filename
])
cmd
.
append
(
os
.
path
.
split
(
cppfilename
)[
-
1
])
cmd
.
append
(
os
.
path
.
split
(
cppfilename
)[
-
1
])
...
...
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