ubuntu12.04配置NVIDIA cuda7.0经验帖

今天安装了最新版本的cuda7.0，第一次安装，并测试成功。先将安装步骤，整理如下。

————————————————————————————————————————————————————————————

配置手册：http://docs.nvidia.com/cuda/cuda-getting-started-guide-for-linux/index.html

这个针对linux系统的，讲的很复杂，看了一遍没看懂，最后参考Rachel-Zhang的安装文档（http://blog.csdn.net/abcjennifer/article/details/23016583），安装成功。

——————————————————————————————

一、 判断安装环境：

The setup of CUDA development tools on a system running the appropriate version of Linux consists of a few simple steps:

• Verify the system has a CUDA-capable GPU.
• Verify the system has a supported version of Linux.
• Verify the system has gcc installed

主要检查是否有gpu，是否安装了gcc

二、下载镜像：

        https://developer.nvidia.com/cuda-downloads 

       下.run好了，有点大1.1G，可到我的云盘里下载。

三、Ctrl+Alt+F1, 登陆

1 删除已有nvidia，拉黑名单

1. sudo apt-get remove --purge nvidia*  
2. sudo nano /etc/modprobe.d/blacklist.conf  

blacklist.conf加入

[plain] view plaincopy

1. lacklist amd76x_edac  
2. blacklist vga16fb  
3. blacklist nouveau  
4. blacklist rivafb  
5. blacklist nvidiafb  
6. blacklist rivatv  

2 安装

[plain] view plaincopy

1. sudo service lightdm stop  
2. chmod +x cuda_7.0.79_linux_64.run  
3. sudo ./cuda_7.0.79_linux_64.run  

Cuda7.0集成了toolkit，sdk和driver, 所以安装的时候一路同意下来就好了

得到以下安装结果

 Driver:   Installed
   Toolkit:  Installed In /Usr/Local/Cuda-7.0
   Samples:  Installation Failed. Missing Required Libraries.

也就是SDK安装失败，这个官方手册上也提到了解决方案。

3 此时可以单独安装sdk

解压sdk：

[plain] view plaincopy

1. sh cuda_7.0.79_linux_64.run -extract=/path/to/extract/dir/  

到根目录下的/path/to/extract/dir/去找cuda-samples_7.0.79_*******.run, 安装

[plain] view plaincopy

1. sh cuda-samples_7.0.79_*******.run  

4 Path配置：

$sudo gedit etc/profile或~/.bashrc

[plain] view plaincopy

1. export PATH=$PATH:/usr/local/cuda-5.5/bin  
2. export LD_LIBRARY_PATH=/usr/local/cuda-5.5/lib64:/lib  

使之生效：

source /etc/profile（对应profile）或sudo ldconfig（对应bashrc）

5 然后验证， nvcc -V

6 测试

Hello World
In the spirit of all good programming assignments, we’ll start off with a simple “Hello World” app that will give a good starting place for making your own programs.
Create a new folder in your CUDA work directory and open a new file:

$ mkdir ~/Documents/CUDA/CUDA_Hello_World
$ cd ~/Documents/CUDA/CUDA_Hello_World
$ nano CUDA_Hello_World.cu

In that file, write out the basic Hello World program:

#include <stdio.h>
__global__ void kernel(void) {
}
int main(void) {
    kernel<<<1,1>>>();
    printf("Hello, World!\n");
    return 0;
}

Save and exit the editor. Then, compile the program:

$ nvcc CUDA_Hello_World.cu -o CUDA_Hello_World

There shouldn’t be any errors. Therefore, you can just run it!

$ ./CUDA_Hello_World