Getting Started with GPU Day

This short guide will walk you through launching a Jupyter server on the reserved multi-instance GPUs (MIG) slices during GPU Day.

Reservation Details

• Reservation name: GPU_DAY

• Reservation time: April 14, 2025, 8:00 AM - 6:00 PM

• GPU Type: 10GiB MIG slices

Launch a Jupyter Server

Log in to the Web Portal

Navigate to http://sol.asu.edu and log in using your ASURITE and password. You may be prompted to authenticate using DUO.

Request a Jupyter Server

1. Click on Interactive Apps → Jupyter.

2. Complete the form with the following information:

   1. Account: use the default account.

   2. Partition: htc.

   3. QOS: public.

   4. CPU Core Allocation: 1.

   5. Memory Allocation: 10.

   6. GPU Resources: gpu:1g.10gb:1.

   7. Jupyter Wall Time: 0-4.

   8. Additional sbatch options: --reservation=GPU_DAY.

3. Click “Launch”.

4. Once the session is running, click “Connect to Jupyter.”

3. Start a new Jupyter Notebook

From the Notebooks available on the Launcher tab, select the rapids25.02 option.

Access to the Bootcamp’s Jupyter Notebooks

To access the bootcamp’s Jupyter Notebooks, follow these instructions:

1. Open a terminal instance from the Launcher or click File → New → Terminal.

2. Copy the notebooks cp -r /data/sse/notebooks/python_notebooks .

3. You may close the terminal instance now.

4. Open the notebooks using the left-hand side file navigation menu.