Install NVIDIA driver + CUDA + MATLAB in Ubuntu 14.04

Install NVIDIA driver + CUDA + MATLAB in Ubuntu 14.04

Ubuntu14.04 install NVIDIA driver + CUDA + MATLAB

1. Install the NVIDIA graphics card driver

1. The nouveau error message is displayed when the video card driver is installed. You need to uninstall this module to continue.

2. In the/etc/modprobe. d/blacklist. conf file, add the nouveau module. Use the nano command to open and edit files (when vi and vim are unavailable)

3. If the nouveau module still exists after the blacklist is added, you can directly remove the kernel module. The command is:

$ Sudo mv/lib/modules/3.0.0-12-generic/kernel/drivers/gpu/drm/nouveau. ko/lib/modules/3.0.0-12-generic/kernel/drivers/gpu/drm/nouveau/nouveau.ko.org

$ Sudo update-initramfs-u Note: Use the tab key to complete the command line. Specific Environment Analysis

4. Don't forget to restart the system to determine whether the nouveau module is completely disabled.

5 ../driver name: Install the nvidia display driver. An error is reported, prompting you to switch to the no X server running status for installation. Run the command $ sudo service lightdm stop (different from the Red Hat and CentOS systems) to stop running ubuntu graphics desktop)

6. After the installation is complete, run the following command to check whether the GPU card has been recognized by the system:

$ Sudo cat/proc/driver/nvidia/gpus /......

Ii. Install the cuda Development Environment

1./cuda file name. Install it step by step according to the software prompts. But there is something worth attention.

When all the installation procedures are complete, an error will be reported, prompting that Samples fails to be installed. That is to say, the SDK fails to be installed. The solution is to install the SDK separately. The method is as follows:

First, extract the SDK installation file from the cuda file:

./Cuda_5.0_linux_64.run-extract =/path/to/extract/dir/the first is the cuda installation source file, followed by the target path to be extracted, which can be customized

To the directory/path/to/extract/dir/find the cuda-samples_5.0 _ *******. run the installation.

After the SDK is installed, all required components are successfully installed.

2. Configure the cuda environment variable and run $ sudo gedit ~ /. Bashrc. Add the following two lines to the file:

Export PATH = $ PATH:/usr/local/cuda-5.0/bin

Export LD_LIBRARY_PATH =/usr/local/cuda-5.0/lib64:/lib save and exit, then execute

Sudo ldconfig

3. Verify that the cuda configuration is successful. You can run a simple command $ nvcc-V. If the configuration is successful, the output content will appear:

3. Install the matlan 2013a for linux environment

1. Copy and install the ISO file (the company NAS has the files required for the downloaded installation) to ubuntu.

There is a. lic file in the Serial folder, which is used to activate the software after installation, so you also need to copy it to ubuntu.

2. Because the ISO file is too large (around GB) to be engraved, you can directly mount the ISO file to read it. Run the command: mount-o loop maid macunix.iso/home/wangbo/followed by the target path to be mounted

3. cd to/home/wangbo/and run the install file to start installation. Note that the above operations must be performed on the command terminal on the graphic desktop, because when we run. /install command, will automatically pop up a matlab installation graphical interface, if it is in F1-F6 non graphical interface mode, execute. /install, an error will be reported, prompting you to be in the X11 display environment, that is, the graphic desktop environment. The newer Ubuntu version will find a depressing problem, that is, right-click the desktop, and the terminal option is not enabled, and the terminal icon cannot be found anywhere on the desktop, if the installation program cannot be opened, you cannot run it. In this case, you don't have to worry about it. Press ctrl + alt + t on the desktop, and a terminal will pop up immediately on the desktop, then we can execute the install installation program on the terminal. After the command is executed, the effect is the same as that when matlab is installed in windows.

4. For the remaining work, refer:

Steps for installing and creating a shortcut for Linux in Ubuntu 14.04

5. Finally, after matlab is installed and activated successfully, restart the system.

6. after the restart, go to/usr/local/Matlab/R2013a/bin/Matlab to start the software. (the actual installation environment is the main path. The environment provided by me is for reference only)

TIPS:

1. Nouveau is an open-source Nvidia graphics card driver. In contrast, Nvidia's official driver is not open-source, and open-source xf86-video-nv features are weak. Compared with the official driver, Nouveau supports the Kernel Mode Setting feature of the new Kernel. In Fedora 12, when the Nouveau driver is used to start the kernel, the optimal resolution of the display is immediately displayed. the startup interface (whether text or graphics) you can use Ctrl + Alt + Fn to completely switch between X and the terminal, greatly enhancing the user experience.

However, the DRM module does not enter the kernel code tree. Yesterday, I saw a message on Phoronix. It was a long time since Linus saw that Noora provided Nouveau and hoped that the DRM of Nouveau would also enter the kernel. Nouveau maintainers and Fedora maintainers explained that the reason is that Nouveau relies on reverse engineering to understand some internal mechanisms of the nv gpu and uses some Microcode, these Microcode-related codes may be copyrighted when they enter the kernel. Linus replied that these reasons are all BS (from several Linus speeches, we can see that this is really cool). Does Fedora have no copyright issue? Yes. The above reason is BS. Otherwise, the Nouveau driver will not be distributed in the future.

Today, I saw the above message. The following is the message: Red Hat company David Airlie and Ben Skeggs modified part of Nouveau's kernel code, microcode is loaded using the Firmware interface of the kernel. This avoids copyright issues.

Next, I will reference the original article:

In this pull request, there is the Nouveau driver that is set to go in the Linux 2.6.33 kernel under the staging area

In 2.6.33, we can use the Nouveau driver that officially enters the kernel to obtain various new features.

PS: although the NV driver does not support KMS, it supports OpenCL (G80 +) and XvMC (G80-) and Vdpau (G80 +) for video acceleration ). Nouveau does not support these features. In fact, the 3D function is not directly supported by Nouveau. The current development trend is that Mesa and Gallium3D provide OpenGL (or even D3D), OpenVG, and OpenCL support. The video card driver only performs basic interaction with the video card. Recently, some Kernel images have become highlights. page flipping ioctl enters 2.6.33 ("It is said that" it is useful for both X and Wayland ), now, the KMS drivers of Intel/ATI/NV are constantly improved, and the graphic experience of Linux users will be enhanced gradually.

2. Installing GCC in Ubuntu is a little different from other Linux systems.
Method 1: This method is super simple:
Sudo apt-get build-depgcc
The above command is enough.

Method 2:

Sudo apt-get install build-essential

It's still simple. A command can also be used.

After installation, run the gcc -- version command to view the version. The output is as follows:

Gcc (GCC) 4.2.3 (Ubuntu4.2.3-2ubuntu7)
Copyright (C) 2007 FreeSoftwareFoundation, Inc.

3. Run the nano command to open the file. Enter the file directly and insert the text. Then, press ctrl + o to save the changes.

Installation Process of CUDA (including GPU driver) in Ubuntu

Caffe + Ubuntu 14.04 64bit + CUDA 6.5 configuration instructions

Install and configure CUDA in Ubuntu 14.04

Install NVIDIA CUDA5.5 in Ubuntu 12.04

Install Theano + CUDA on Ubuntu

For Ubuntu 12.04 CUDA5.5 installation, see the following link Ubuntu 12.04 installation CUDA-5.5