GRIT HPC Setup Guide for GEOG-288KC

This guide provides step-by-step instructions for setting up access to the GRIT HPC system and configuring Jupyter Notebook with the geoAI kernel.

Prerequisites

GRIT user account credentials
SSH client installed on your local machine
Web browser with JavaScript enabled

Setup Instructions

Step 1: Create a GRIT User Account

Contact your course administrator to request a GRIT user account if you don’t already have one.

Step 2: Configure SSH Access and Kernel Symbolic Link

Open your terminal/command prompt
Connect to the GRIT HPC system via SSH:
```
ssh <net_id>@hpc.grit.ucsb.edu
```

Once logged in, create a symbolic link to the geoAI kernel:

ln -s /home/g288kc/.local/share/jupyter/kernels/geoai ~/.local/share/jupyter/kernels/

Exit the SSH session:
```
exit
```

Step 3: Access the GRIT HPC Web Portal

Open your web browser and navigate to: https://hpc.grit.ucsb.edu
Log in using your GRIT credentials

GRIT Login Page

Step 4: Launch Jupyter Notebook

Once logged in, click on “My interactive sessions” in the menu
Under the “Servers” section, select “Jupyter Notebook”

Interactive Sessions Menu

Step 5: Configure Jupyter Notebook Instance

Create a new Jupyter Notebook instance with the following parameters:

Parameter	Value
Username	Your GRIT username
Partition	`grit_nodes`
Job Duration	`168` (maximum allowed, in hours)
CPUs	`4` (adjust based on needs)
RAM	`16` (GB, adjust based on needs)
Job Name	Optional - leave empty or provide descriptive name

Click “Launch” to start your Jupyter Notebook instance.

Step 6: Configure the geoAI Kernel

To enable AI features in your Jupyter Notebook:

Once your Jupyter Notebook is running, go to the “Kernel” menu
Select “Change Kernel…”

Kernel Menu
From the list of available kernels, select “geoAI Course”

Kernel Selection

Important Notes

Session Duration: The maximum job duration is 168 hours (7 days). Plan your work accordingly.
Resource Allocation: The CPU and RAM values provided are recommendations. Adjust based on your computational needs.
Kernel Link: The symbolic link created in Step 2 is essential for accessing the geoAI kernel. Without it, the kernel won’t appear in your Jupyter options.

Troubleshooting

Cannot see geoAI kernel

Verify the symbolic link was created correctly by running:
```
ls -la ~/.local/share/jupyter/kernels/
```
Restart your Jupyter Notebook session

SSH Connection Failed

Verify you’re using the correct hostname: hpc.grit.ucsb.edu
Check that your GRIT account is active
Ensure you’re connected to the internet

Jupyter Session Won’t Start

Check if you have any existing sessions that need to be terminated
Verify you’ve selected the correct partition (grit_nodes)
Try reducing the requested resources (CPUs/RAM)

Support

For additional help:

Contact your course instructor
Reach out to GRIT HPC support
Consult the UCSB HPC documentation

Last updated: September 2025 Based on instructions from tjaartvdwalt/grit_user_setup.md

--- title: "HPC Setup Notes" subtitle: "GRIT HPC Notes for GEOG-288KC" jupyter: geoai format: html: code-fold: false --- # GRIT HPC Setup Guide for GEOG-288KC This guide provides step-by-step instructions for setting up access to the GRIT HPC system and configuring Jupyter Notebook with the geoAI kernel. ## Prerequisites - GRIT user account credentials - SSH client installed on your local machine - Web browser with JavaScript enabled ## Setup Instructions ### Step 1: Create a GRIT User Account Contact your course administrator to request a GRIT user account if you don't already have one. ### Step 2: Configure SSH Access and Kernel Symbolic Link 1. Open your terminal/command prompt 2. Connect to the GRIT HPC system via SSH: ```bash ssh <net_id>@hpc.grit.ucsb.edu ``` 3. Once logged in, create a symbolic link to the geoAI kernel: ```bash ln -s /home/g288kc/.local/share/jupyter/kernels/geoai ~/.local/share/jupyter/kernels/ ``` 4. Exit the SSH session: ```bash exit ``` ### Step 3: Access the GRIT HPC Web Portal 1. Open your web browser and navigate to: <https://hpc.grit.ucsb.edu> 2. Log in using your GRIT credentials ![GRIT Login Page](https://gist.github.com/user-attachments/assets/5c749926-9a50-463d-82c0-3bff1f05a529) ### Step 4: Launch Jupyter Notebook 1. Once logged in, click on **"My interactive sessions"** in the menu 2. Under the **"Servers"** section, select **"Jupyter Notebook"** ![Interactive Sessions Menu](https://gist.github.com/user-attachments/assets/ffb16d7f-e77c-4c7f-a1e1-7262578e58b7) ### Step 5: Configure Jupyter Notebook Instance Create a new Jupyter Notebook instance with the following parameters: | Parameter | Value | |-----------|-------| | **Username** | Your GRIT username | | **Partition** | `grit_nodes` | | **Job Duration** | `168` (maximum allowed, in hours) | | **CPUs** | `4` (adjust based on needs) | | **RAM** | `16` (GB, adjust based on needs) | | **Job Name** | Optional - leave empty or provide descriptive name | Click **"Launch"** to start your Jupyter Notebook instance. ### Step 6: Configure the geoAI Kernel To enable AI features in your Jupyter Notebook: 1. Once your Jupyter Notebook is running, go to the **"Kernel"** menu 2. Select **"Change Kernel..."** ![Kernel Menu](https://gist.github.com/user-attachments/assets/9e2e32a6-7e4d-4b28-9520-c244b11c4656) 3. From the list of available kernels, select **"geoAI Course"** ![Kernel Selection](https://gist.github.com/user-attachments/assets/d9d5cf04-2e70-4ecf-bc44-0f2e17a2c8a3) ## Important Notes - **Session Duration**: The maximum job duration is 168 hours (7 days). Plan your work accordingly. - **Resource Allocation**: The CPU and RAM values provided are recommendations. Adjust based on your computational needs. - **Kernel Link**: The symbolic link created in Step 2 is essential for accessing the geoAI kernel. Without it, the kernel won't appear in your Jupyter options. ## Troubleshooting ### Cannot see geoAI kernel - Verify the symbolic link was created correctly by running: ```bash ls -la ~/.local/share/jupyter/kernels/ ``` - Restart your Jupyter Notebook session ### SSH Connection Failed - Verify you're using the correct hostname: `hpc.grit.ucsb.edu` - Check that your GRIT account is active - Ensure you're connected to the internet ### Jupyter Session Won't Start - Check if you have any existing sessions that need to be terminated - Verify you've selected the correct partition (`grit_nodes`) - Try reducing the requested resources (CPUs/RAM) ## Support For additional help: - Contact your course instructor - Reach out to GRIT HPC support - Consult the [UCSB HPC documentation](https://csc.cnsi.ucsb.edu/) --- *Last updated: September 2025* *Based on instructions from [tjaartvdwalt/grit_user_setup.md](https://gist.github.com/tjaartvdwalt/e4080d91aa38d0758d8ef813f75d64ac)*