zoukankan      html  css  js  c++  java
  • Nvidia CUDA 6 Installed In Ubuntu 12.04

    环境:ubuntu 12.04 (x64)

    如果不能够 service lightdm stop,显示:unknown service 或者其他的 

    sudo /etc/init.d/lightdm restart

    cuda 6.0

    1)Verify You Have a CUDA-Capable GPU

    To verify that your GPU is CUDA-capable, go to your distribution’s equivalent of SystemProperties, or, from the command line, enter:

    1
    lspci | grep -i nvidia


    2) Verify the System Has gcc Installed

    1
    gcc --version

    一般系统安装完毕后便自带C编译环境和Python编译环境

    3)安装必要的库

    If you wish to build allof the samples, including those with graphical rather than
    command-line interfaces, you may need to install additional system libraries or
    headers if you have not done so before.

    1
    2
    sudo apt-get install freeglut3-dev build-essential libx11-dev libxmu-dev
    libxi-dev libgl1-mesa-glx libglu1-mesa libglu1-mesa-dev


    注:
    安装libgl1-mesa-* library时可能会提示你缺少依赖包,请谨记安装提示信息中Recommended依赖包,否则,呵呵呵…

    4)卸载(禁止)存在的NVIDIA(nouveau)驱动

    一般Ubuntu系统安装完毕后,会弹出Additional Driver警示框,提醒安装驱动,为了避免兼容性问题,如果已经安装的请卸载PS:The CUDA 5.5 installers include the CUDA Toolkit, SDK code samples, Nsight Visual Studio edition (for Windows) and Nsight Eclipse Edition (for Linux / Mac OS X), and developer drivers.

    1
    sudo nvidia-uninstall

    如果卸载失败,或者卸载完成后,建议这样做一下

    1
    sudo apt-get remove --purge nvidia*

    禁止开源驱动nouveau的开机加载

    1
    sudo gedit /etc/modprobe.d/blacklist.conf

    在文件后添加如下内容保存后重启

    1
    2
    blacklist nouveau
    options nouveau modeset=0


    5)Installation

    Exit the GUI if you are in a GUI environment by pressing Ctrl-Alt-F1 and login as root, then type into the commond below and press Enter

    1
    service lightdm stop

    Install the CUDA Toolkit (xx in 5.5.xx is the minor version of the installation package)by running the downloaded .runfile as a superuser.

    1
    sudo sh cuda_5.5.xx_linux_32_rhel5.x.run

    这个初始化时间有点长,耐心等待…

    安装完成后切换回GUI界面

    1
    service lightdm start




    6)Define the environment variables

    我习惯用如下方式进行环境变量的添加修改:

    首先

    1
    gedit ~/.bashrc

    添加如下内容

    1
    2
    export PATH=/usr/local/cuda-6.0/bin:$PATH
    export LD_LIBRARY_PATH=/usr/local/cuda-6.0/lib64:$LD_LIBRARY_PATH

    然后

    1
    gedit /etc/ld.so.conf

    添加如下内容

    1
    /usr/local/cuda-6.0/lib64

    完成之后执行如下命令

    1
    sudo ldconfig

    验证是否添加路径成功

    1
    ldconfig -v|grep cuda

    以上内容其实在你安装完毕CUDA Toolkit之后的信息中都有所提示!文章最后会给出安装完毕后的截图!

    7)关于Bug:

    执行 nvidia.settings 出现错误提示框 no driver detected 以及 nvidia driver is activated but not currently in use

    不知道怎么解决,郁闷中…反正console project都能运行,只是进入Ubuntu系统时因为驱动未安装好的缘故会花屏而已,另外建议运行一下Samples中的1_Utilities中的deviceQuery程序作为测试!

  • 相关阅读:
    idea中代码提交流程(git版)
    《如何做好软件设计》:设计原则
    用基础Array数组实现动态数组、链表、栈和队列
    使用Redis+SpringBoot实现定时任务测试
    分布式锁的三种实现方式
    RabbitMQ镜像队列集群搭建、与SpringBoot整合
    RabbitMQ简介、安装、基本特性API--Java测试
    分布式文件系统FastDFS简介、搭建、与SpringBoot整合实现图片上传
    Elasticsearch--Logstash定时同步MySQL数据到Elasticsearch
    Elasticsearch 分片集群原理、搭建、与SpringBoot整合
  • 原文地址:https://www.cnblogs.com/assassin/p/3689974.html
Copyright © 2011-2022 走看看