Installing NVIDIA Drivers

Introduction

To fully leverage the capabilities of your GPU-accelerated VM within the Ori Global Cloud (OGC) platform, it is essential to install the appropriate NVIDIA drivers. This ensures that your VM can efficiently utilize the underlying hardware for GPU-intensive tasks such as machine learning, deep learning, and high-performance computing. This guide will walk you through the process of installing NVIDIA GPU drivers on your VM.

Prerequisites

Ensure that your VM instance is available.
Verify that you have access to your VM via ssh. Use the Ori provided ssh command on the VM details page.

Installation Steps

Prepare the System:
- Find information about your systems GPU
```
lspci | grep -i nvidia
```
- Check your Linux OS version:
```
cat /etc/os-release
```
- Check your VM CPU architecture version:
```
uname -a
```
Follow the NVIDIA developer install instructions:
- The guide is here, follow the selection process with the information collected above:
  - https://developer.nvidia.com/cuda-downloads?target_os=Linux
- Input the commands provided on the NVIDIA installation guide
Verify the Installation:
- Once the installation is complete, reboot the VM
```
sudo reboot
```
- Verify the driver installation:
```
nvidia-smi
```
- The nvidia-smi utility should display information about the GPU and the installed driver.

Troubleshooting

If you encounter issues during the driver installation, consider the following troubleshooting steps:

Check the NVIDIA developer forums and knowledge base for solutions to common issues.
Ensure that any previous NVIDIA driver installations are completely removed before attempting a fresh installation.

Additional Resources

For more detailed instructions, advanced configurations, and troubleshooting advice, refer to the official NVIDIA documentation available on the NVIDIA developer pages. You can also find support and community discussions that may assist with unique installation scenarios or issues.