This topic describes how to download and install the GPU driver. If you do not configure automatic installation of the GPU driver when you create a GPU instance, you must install it manually after the creation.
Download the GPU driver
- Go to NVIDIA website.
- Find the drivers for your NVIDIA products, and click Search. The following table shows relevant metrics.
Item gn4 gn5 gn5i gn6v vgn5i gn6i Product type: Tesla Tesla Tesla Tesla Tesla Tesla Product series M-Class P-Series P-Series V-Series P-Series T-Series Product M40 Tesla P100 Tesla P4 Tesla V100 Tesla P4 Tesla T4 Operating system (OS) Select a release according to the image of your instance.Note
- If your OS is not in the drop-down list, click Select All Operating Systems at the bottom of the list.
- For Debian, select Linux 64-bit.
- Confirm the information is correct, then click Download.
Install the GPU driver for a Linux instance
- Download and install the kernel-devel and kernel-header packages based on your kernel version.
Note If version conflict exists between kernel-devel and your kernel, driver compilation errors will occur during driver rpm installation. You can run rpm –qa | grep kernel in your instance to check version consistency before installing the driver.
- Run the sudo rpm -qa | grep $(uname -r) command to verify that the kernel-devel and kernel-header packages are successfully installed.
For example, if you are using CentOS 7.3 and the following information appears, this means that the installation is successfully completed.
kernel-3.10.0-514.26.2.el7.x86_64 kernel-headers-3.10.0-514.26.2.el7.x86_64 kernel-tools-libs-3.10.0-514.26.2.el7.x86_64 python-perf-3.10.0-514.26.2.el7.x86_64 kernel-tools-3.10.0-514.26.2.el7.x86_64
- Install the GPU driver according to the description in the ADDITIONAL INFORMATION area of the GPU download page.
Take Linux 64-bit Ubuntu 14.04 for example: