Getting Started with GPU-Powered JupyterLab on Brev🤙

Follow these steps to launch your GPU-enabled JupyterLab environment using Brev.

1. Allow the Instance to Start

Wait for your Brev instance to finish building before proceeding.

2. Install Brev CLI (If Not Already Installed)

Copy and paste the “Install the CLI” command from your Brev instance page into your terminal.

3. Connect to Your Instance

Under the “Open a terminal locally” section, copy your brev shell <instance-name> command and run it in your terminal.

4. Change into Proper Directory

Run the following command in your terminal: cd gpu-python/

5. Start the Docker Container

Run the following command in your terminal: docker compose up

6. Refresh Brev Page

Refresh the Brev page, open your notebook, and enjoy!

Debugging

If you try to start Jupyter but get an error like:

Error: [Errno 98] Address already in use It means another process is already using port 8888. This is usually caused by a Jupyter service that’s running automatically in the background.

Follow these steps to diagnose and fix the issue:

1. Check What’s Using Port 8888

Run: lsof -i :8888 Example output: COMMAND PID USER FD TYPE DEVICE SIZE/OFF NODE NAME jupyter-lab 29753 ubuntu 6u IPv4 129739 0t0 TCP *:8888 (LISTEN)

2. Stop and Disable Auto-Start

If Jupyter is being started automatically by systemd, stop it with: sudo systemctl stop jupyter sudo systemctl disable jupyter

3. Verify It’s Gone

lsof -i :8888 If nothing shows up, the port is now free.

Accelerated Python User Guide

Notebooks

Notebook	Link
Chapter 1: GPU Computing Basics
Chapter 2: Brief Intro to CUDA
Chapter 3: Python on the GPU
Chapter 4: Scientific Computing with CuPy
Chapter 5: CUDA Kernels with Numba
Chapter 6: Intro to nvmath-python
Chapter 7: Intro to cuDF
Chapter 8: Intro to cuML
Chapter 9: Intro to cuGraph
Chapter 10: Developer Tools
Chapter 11: Distributed Computing with cuPyNumeric
Chapter 12: Intro to NVIDIA Warp