Installation tutorial for Nvidia Tesla K80 graphics driver under ubuntu14.04

Source: Internet
Author: User

How can deep learning not deal with graphics cards that are "nuclear"?

The rise of artificial intelligence, in addition to the massive increase in data volume, the continuous improvement of algorithms, the gradual improvement of computing power, but also inseparable from the gradual improvement of software infrastructure. The current mainstream deep learning tool software, whether Caffe or Theano or TensorFlow, is all about the support of GPU graphics acceleration technology. The current mainstream GPU-accelerated deep learning environment is the CUDA under Linux. As an excellent product of Nvidia's current computational acceleration, Tesla K80 has become an ideal target. However, due to the Linux open Source feature, Nouveau becomes the default Nvidia graphics driver in a series of Linux operating systems such as Ubuntu. However as nvidia such a high-cold company, market-style products are not recognized. So in order to protect their graphics card performance, NVIDIA graphics card has its own driver, can be the root of the red and grass-roots version of the drive between the existence of an amicable conflict. So want to use Ubuntu operating system to drive K80 still have to spend some trouble, but the core of the problem is to solve the two-class drive between the opposing relationship, the method is very simple, leather off a good. Here is a detailed procedure.

1, the installation of ubuntu14.04 (the specific mode of operation is not much to say, but to emphasize the point is that Dell's personal pc in the installation of Ubuntu may often have a graphical interface is not available, the problem is also estimated to be incompatible with the driver, if there is no independent graphics card, or do not toss. Fortunately, this build environment is PowerEdgeC4230 this Dell's server, so it is not a big problem to install a gnome)

2, because it is the conflict of the drive, then nature is to kill and the graphics card combination is not so close to the grassroots board drive Nouveau, to join the blacklist so that we have to do the first thing:

2.1 Create a blacklist-nouveau.conf file under the /etc/modprobe.d/ folder

2.2 Use VI to easily create a new file to write in the file:

      Blacklist nouveau

Options Nouveau modeset=0

Save the file (preferably using the VI editor, try not to use the nano of this less mainstream editor, the previous installation failed perhaps because of the use of Nano editor)

2.3 Enter update initramfs-u command to get updated blacklist

2.4 Enter reboot command to restart the system, strong guarantee blacklist effective.

2.5 Input lsmod | grep nouveau See if Vouveau really got killed, if nothing appears, then congratulations, successfully removed the grass-driven version.

3, download nvidia official K80 Graphics driver, we use the 361 version here. Move the drive files to the/root/folder.

4. If the file is a compressed file, type the unzip command to unzip it and get the. Run File

5, Next is the key step, is also the problem that bothers us all day and night, type the following command:

    ./xxx-nvidia.run--no-opengl-files (important thing to say three times, there is no space between the letter and the letters, there is no space between the letters, there is no space between the '-' and the letters )

6, there is a blue background interface, if it appears (X server is running phenomenon, to note the user state input sudo service LIGHTDM stop Shutdown Desktop manager), and then the Accept protocol, followed by the distribution Provided pre-install scripts failed hints, ignores it, and then continues to install, all the way OK then reboot the system and eventually get the complete GNOME desktop system.

7, check whether the installation is successful, in the command line interface input NVIDIA-SMI test whether the installation is successful, if successful will appear as shown in the situation .

Installation tutorial for Nvidia Tesla K80 graphics driver under ubuntu14.04

Contact Us

The content source of this page is from Internet, which doesn't represent Alibaba Cloud's opinion; products and services mentioned on that page don't have any relationship with Alibaba Cloud. If the content of the page makes you feel confusing, please write us an email, we will handle the problem within 5 days after receiving your email.

If you find any instances of plagiarism from the community, please send an email to: info-contact@alibabacloud.com and provide relevant evidence. A staff member will contact you within 5 working days.

A Free Trial That Lets You Build Big!

Start building with 50+ products and up to 12 months usage for Elastic Compute Service

  • Sales Support

    1 on 1 presale consultation

  • After-Sales Support

    24/7 Technical Support 6 Free Tickets per Quarter Faster Response

  • Alibaba Cloud offers highly flexible support services tailored to meet your exact needs.