cudamat在ubuntu上安装nvcc失败-Java 学习之路

我在EC2上运行了GPU实例，我正在尝试安装cudamat .

我已将副本下载到我的主目录：

git clone https://github.com/cudamat/cudamat

然后我进入该目录并运行：sudo pip install .

错误是“命令'nvcc'因退出状态1而失败 .

下面是完整的消息：解压缩/ home / ubuntu / cudamat从file：/// home / ubuntu / cudamat PATH = / usr运行setup.py（路径：/tmp/pip-s7db_8yv-build/setup.py）egg_info包/ local / sbin：/ usr / local / bin：/ usr / sbin：/ usr / bin：/ sbin：/ bin / tmp / pip-s7db_8yv-build

Installing collected packages: cudamat
  Running setup.py install for cudamat
    PATH = /usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin
/tmp/pip-s7db_8yv-build

    building 'cudamat.libcudamat' extension
    nvcc -I/usr/include/python3.4m -c cudamat/cudamat.cu -o build/temp.linux-x86_64-3.4/cudamat/cudamat.o -O --ptxas-options=-v  --compiler-options=-fPIC
    error: command 'nvcc' failed with exit status 1
    Complete output from command /usr/bin/python3 -c "import setuptools, tokenize;__file__='/tmp/pip-s7db_8yv-build/setup.py';exec(compile(getattr(tokenize, 'open', open)(__file__).read().replace('\r\n', '\n'), __file__, 'exec'))" install --record /tmp/pip-nulrhjnl-record/install-record.txt --single-version-externally-managed --compile:
    PATH = /usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin

/tmp/pip-s7db_8yv-build

running install

running build

running build_py

creating build

creating build/lib.linux-x86_64-3.4

creating build/lib.linux-x86_64-3.4/cudamat

copying cudamat/learn.py -> build/lib.linux-x86_64-3.4/cudamat

copying cudamat/__init__.py -> build/lib.linux-x86_64-3.4/cudamat

copying cudamat/cudamat.py -> build/lib.linux-x86_64-3.4/cudamat

running egg_info

creating cudamat.egg-info

writing cudamat.egg-info/PKG-INFO

writing dependency_links to cudamat.egg-info/dependency_links.txt

writing top-level names to cudamat.egg-info/top_level.txt

writing manifest file 'cudamat.egg-info/SOURCES.txt'

warning: manifest_maker: standard file '-c' not found



reading manifest file 'cudamat.egg-info/SOURCES.txt'

writing manifest file 'cudamat.egg-info/SOURCES.txt'

copying cudamat/cudamat.cu -> build/lib.linux-x86_64-3.4/cudamat

copying cudamat/cudamat_kernels.cu -> build/lib.linux-x86_64-3.4/cudamat

copying cudamat/learn.cu -> build/lib.linux-x86_64-3.4/cudamat

copying cudamat/learn_kernels.cu -> build/lib.linux-x86_64-3.4/cudamat

copying cudamat/rnd_multipliers_32bit.txt -> build/lib.linux-x86_64-3.4/cudamat

running build_ext

building 'cudamat.libcudamat' extension

    creating build/temp.linux-x86_64-3.4

    creating build/temp.linux-x86_64-3.4/cudamat

    nvcc -I/usr/include/python3.4m -c cudamat/cudamat.cu -o build/temp.linux-x86_64-3.4/cudamat/cudamat.o -O --ptxas-options=-v --compiler-options=-fPIC

    error: command 'nvcc' failed with exit status 1

路径有问题 . 因为找不到nvcc . 当我输入时：

echo $PATH
/usr/local/cuda/bin:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games

有人有想法吗？

干杯

2 回答

0
只有在使用sudo时才会出现错误 .

sudo更改$ PATH变量，具体取决于安全策略 .

如果使用以下内容，问题将得到解决：
```
sudo env "PATH=$PATH" pip3 install .
```
回复于 2024-05-03T23:46:10+08:00
0
nvcc没有配置为sudo访问（我可能错误地说这句话，但简而言之，nvcc不适用于sudo）

这个问题有两个解决方案：
- 给suv访问nvcc（你提到过），这可能是效率低下的方法 .
- 安装cudamat时不要使用sudo，而是在cudamat目录中使用以下命令 .
python setup.py install --user

注意：
- 在安装后进行测试时，请不要使用cudamat目录中的 import cudamat ，这将通过错误说明找不到cudamat / cudamat .
- 它将在本地目录中查找 .
- 回家并使用 import cudamat 它会起作用
回复于 2024-05-03T23:46:10+08:00

cudamat在ubuntu上安装nvcc失败

2 回答

相关问题