Installing NVIDIA CUDA Toolkit on Ubuntu 22.04 Cloud GPU Server

Log in to Shape.host Dashboard: Navigate to the Shape.host website and log in to your account.
Create a New Instance:

Click on the “Create” button located at the top right corner of the dashboard.
From the dropdown menu, select “Instances”.

Shape.host Resources dashboard showing the Create menu and instance rebuild controls.

Select Instance Location:

Choose the desired location for your server. For this tutorial, we’ll select “New York, USA”.

Shape.host Create New Instance form with the available data center locations.

Choose a Plan:

Select a plan that fits your requirements. For example, you might choose a plan with 2 cores CPU, 2 GB Memory, and 50 GB SSD disk space.

Shape.host instance form with a server plan and Ubuntu image selected.

Select an Operating System:

Scroll down to the “Choose an image” section and select “Ubuntu 22.04”.

Configure Additional Options:

(Optional) You can configure additional options like User Data Configuration and IPv6 Networking.
Enter a hostname for your instance, e.g., “Tutorial Ubuntu”.
Click on the “Create instance” button to deploy the instance.

Shape.host instance form showing hostname, instance count and the Create Instance button.

Step 2: Connecting to Your Instance

Once your instance is created, you need to connect to it using SSH.

Retrieve SSH Credentials:

Note the IP address of your newly created instance from the Shape.host dashboard.

Connect via SSH:

Open a terminal on your local machine.
Use the following command to connect to your instance:
sh ssh root@your_instance_ip
Replace your_instance_ip with the actual IP address of your instance.

Step 3: Update the System

Before installing CUDA, make sure your system is updated.

sudo apt update
sudo apt upgrade -y

Terminal showing APT package management commands and output.

Step 4: Install NVIDIA CUDA Toolkit

Add NVIDIA Package Repository:

   sudo apt-get update
   sudo apt-get install -y wget
   wget https://developer.download.nvidia.com/compute/cuda/repos/ubuntu2204/x86_64/cuda-ubuntu2204.pin
   sudo mv cuda-ubuntu2204.pin /etc/apt/preferences.d/cuda-repository-pin-600

Terminal showing systemd service management commands or output.

Add the NVIDIA GPG Key:

   sudo apt-key adv --fetch-keys https://developer.download.nvidia.com/compute/cuda/repos/ubuntu2204/x86_64/3bf863cc.pub

Add the CUDA Repository:

   sudo add-apt-repository "deb https://developer.download.nvidia.com/compute/cuda/repos/ubuntu2204/x86_64/ /"

Update the Package Lists:

   sudo apt update

Install CUDA Toolkit:

   sudo apt install -y cuda

Set Up Environment Variables:
Add the following lines to your ~/.bashrc file to set up the environment variables:

   export PATH=/usr/local/cuda-11.7/bin${PATH:+:${PATH}}
   export LD_LIBRARY_PATH=/usr/local/cuda-11.7/lib64${LD_LIBRARY_PATH:+:${LD_LIBRARY_PATH}}

Source the ~/.bashrc file to apply the changes:

   source ~/.bashrc

Verify Installation:
To verify that the CUDA toolkit is installed correctly, run:

   nvcc -V

This command should display the version of CUDA installed on your system.

Shape.host offers robust and flexible Linux SSD VPS services, making it an excellent choice for hosting applications like NVIDIA CUDA Toolkit. With their user-friendly dashboard, easy instance deployment, and a variety of plans to choose from, Shape.host provides the tools you need to set up your cloud infrastructure quickly and efficiently. Whether you are setting up a small team chat server or deploying a larger application, Shape.host’s reliable services ensure that your projects run smoothly and securely.