/google-cloud-gpu-availability-script

Brute forte to setup GPUs on Google Cloud compute instances for colab enterprise, jupyter notebooks, and kubernetes clusters

Primary LanguageShell

Brute force to setup GPU on Google Cloud / GCP

Script to set up a Google Cloud Platform (GCP) instance with an L4 GPU and n1-standard-8 machine type. It also installs Jupyter Notebook and makes it accessible remotely.

What You'll Need

  1. A GCP account and project.
  2. Google Cloud SDK installed and configured.

Getting Started

  1. Clone this repo:

    git clone https://github.com/yourusername/gcp-instance-setup.git
cd gcp-instance-setup

  2. Make the script executable:

    chmod +x setup_gcp_instance.sh

  3. Update your project ID:

    Open setup_gcp_instance.sh and replace YOUR_PROJECT_ID with your actual GCP project ID.

  4. Run the script:

    ./setup_gcp_instance.sh

  5. Follow the instructions:

    The script will tell you how to access the Jupyter Notebook once it's set up.

What the Script Does

  1. Sets your GCP project:

    gcloud config set project YOUR_PROJECT_ID

  2. Gets available zones:

    gcloud compute zones list --format="value(name)"

  3. Tries to create an instance in each zone: It loops through zones until it finds one that works.

  4. Installs Jupyter Notebook: SSHs into the instance, installs Jupyter, and starts it.

  5. Provides access instructions: Tells you how to SSH into the instance and access Jupyter Notebook.

Notes

  • Make sure you have enough quota for L4 GPUs and n1-standard-8 instances.
  • The instance is set to terminate during maintenance. You can change this in the script if needed.

License

This project is licensed under the MIT License. See the LICENSE file for details.

Thanks

Happy computing!