Skip to main content

FluidStack

What is FluidStack​

FluidStack Unlimited scale AI & LLM training with FluidStack. Access 50000+ GPUs including H100s and A100s from a global data center network.

How to register and create a new instance in Lambda Cloud​

The process to start using FluidStack includes a series of straightforward steps.

Create a New Instance​

  1. Start Instance Creation:
  • Click on "Virtual Machine" in the left side of the dashboard to begin setting up a new instance.
  1. Select OS Template:
  • Choose "Ubuntu 22.04 (Plain)" as the operating system template for a clean and minimal setup.
  1. Select Instance Type:
  • Opt for a GPU instance, specifically the A100 PCIE 80GB, which offers robust performance for demanding applications.
  1. Set Up Credentials:
  • If you have not yet set up an SSH key, select "add a new SSH key" and follow the provided steps to create and register a key.
  1. Deploy the Instance:
  • Click "Deploy" to initiate the creation of your instance.
  1. Monitor Instance Setup:
  • Wait for the instance to be fully operational and note down the IP address once it's displayed.

Connect to Your Instance​

  1. Initialize and Connect to Your Instance:
  • Ensure that your instance is active and the IP address is ready.
  • To establish a secure SSH connection, open your terminal and enter:
ssh ubuntu@<your-ip>
  • This command initiates a secure shell connection to your instance, allowing you to start managing it remotely.

Connect your FluidStack Instance to LooPIN Network​

Prerequisites​

Before you get started, ensure you meet the following requirements:

  • A stable network connection and open ports for cluster access.
  • An Ubuntu system with Nvidia GPUs installed, as currently, LooPIN only supports this configuration.

Steps to Join LooPIN​

  1. Register on LooPIN: Head over to loopin.network and sign up using your email or Google account.
  2. Prepare Your Machine: Ensure you have a bare metal machine with an Nvidia GPU. Navigate to the "server" section on the navbar and click on "Deploy a new machine".
  3. CLI Worker Registration:
  • Open a terminal on your Ubuntu system.
  • Run the following command to register your machine as a LooPIN worker (Find it in your guidance page):
curl -fsSL https://files.loopin.network/install.sh | sudo bash -s -- --code <your-unique-code>
  • This command downloads and executes the LooPIN installation script, registering your machine with the provided unique code.
  1. Installation and Verification: The script will check if your machine meets the requirements and will install necessary drivers, including the Nvidia GPU driver. It will then verify your setup and register your worker. Upon successful installation, you should see the message:
LooPIN Client Connected! Waiting for the on-chain verification...

This indicates that the client application is installed and awaiting connection. The first-time connection may take approximately 5-10 minutes, depending on your network speed.

  1. Join the Cluster: Once your machine connects to our cluster and shows an online tag, you are ready to contribute your idle GPU hours to the pool or engage in liquidity swapping.

What You'll Need​

For a smooth setup, ensure you have:

  • A stable internet connection to download necessary files and communicate with the LooPIN network.
  • An Ubuntu system with Nvidia GPUs: This is crucial for the hardware compatibility with LooPIN's requirements.

By following these steps, you'll be fully set to start leveraging your hardware on the LooPIN network.

Go to Liquidity Tutorial and Swap Tutorial for more.