Installation

This section provides step-by-step instructions to install the GRID Enterprise on your system. The GRID Enterprise is designed to work seamlessly on systems with the following requirements.

Requirements

Before proceeding with the installation, ensure that your system meets the following requirements:

  • Operating System: Ubuntu 22.04 / 24.04 The GRID Enterprise is optimized for Ubuntu 22.04 and 24.04, ensuring compatibility with the latest Linux distributions. While it may work on other versions or distributions, these are the officially supported versions.

  • NVIDIA GPU + Drivers: To leverage GRID’s simulation and AI models, an NVIDIA GPU is required, with a VRAM of at least 8 GB (ideally >=16GB). Please ensure the latest NVIDIA drivers (535+) are also installed.

  • Python 3.11+: The GRID Enterprise requires Python 3.11 or higher. We recommend installing through Miniconda or an equivalent environment manager.

  • Docker: Docker is used to containerize applications, ensuring that the GRID Enterprise runs consistently across different environments. Make sure Docker is installed and running on your system. Please find more info here.

  • NVIDIA Container Runtime: The NVIDIA container runtime is required to run GPU-accelerated Docker containers. This runtime allows Docker to utilize the GPU, enabling efficient execution of AI models within containers. Please find more info on the NVIDIA docs page.

Installation Steps

Once your system meets the prerequisites, you can install the GRID Enterprise console with the following command inside your conda/virtual environment:

pip install sf-grid

Note

Please note that all of our setup is based on Linux systems. Some of our features might not work as expected on Windows, WSL or other virtualized environments.

Now that we have all the requirements in place, let's move on to the next section to learn more about initializing and interacting with the GRID Enterprise CLI.