Using cloud GPUs: Difference between revisions

Using cloud GPUs (view source)

Revision as of 15:44, 9 March 2020

299 bytes removed , 4 years ago

no edit summary

Bott

cc_staff

247

edits

@@ Line 77: / Line 77: @@
 <!--T:5-->
 == Preparing a Centos 7 Instance ==
-CentOS has no officialrepository for the Nvidia driver and the kernel module needs to be compile from the vendors sources.
-* Update the OS on the instance to the latest version and reboot the system
+Nvidia provides a rpm package for RHel7, which is compatible with CentOS7.
-<pre>
+https://www.nvidia.com/content/DriverDownload-March2009/confirmation.php?url=/tesla/396.37/nvidia-diag-driver-local-repo-rhel7-396.37-1.0-1.x86_64.rpm&lang=us&type=Tesla
-root@gpu2:~# yum -y update && reboot
-</pre>
-* Installation for the required tools and libraries.
-<pre>
-root@gpu2:~# yum -y install epel-release gcc kernel-devel kernel-headers pciutils
-root@gpu2:~# yum -y dkms
-</pre>
-* Download the required driver source from the Nvidia website and compile the driver and inserted into the kernel.
+* Download and install the vendors rpm, which sets up the repository files to use yum for the installation.
-(https://www.nvidia.com/content/DriverDownload-March2009/confirmation.php?url=/tesla/440.33.01/NVIDIA-Linux-x86_64-440.33.01.run&lang=us&type=Tesla)
 <pre>
-wget http://us.download.nvidia.com/tesla/440.33.01/NVIDIA-Linux-x86_64-440.33.01.run && chmod 755 NVIDIA-Linux-x86_64-440.33.01.run && ./NVIDIA-Linux-x86_64-440.33.01.run
+[root@testgpu-03 centos]# wget http://us.download.nvidia.com/tesla/396.37/nvidia-diag-driver-local-repo-rhel7-396.37-1.0-1.x86_64.rpm && rpm -ivh nvidia-diag-driver-local-repo-rhel7-396.37-1.0-1.x86_64.rpm
 </pre>
 </translate>