Open OnDemand is a tool that allows users to launch Graphical User Interfaces (GUIs) based applications are accessible without modifying your HPC environment. You can log into the Open OnDemand interface at https://ood.hpc.nyu.edu. Once logged in, select the Interactive Apps menu, select the desired application, and submit the job based on required resources and options. 

OOD + Singularity + conda

This page describes how to use your Singularity with conda environment in Open OnDemand (OOD) GUI at Greene. 

Log Into Greene via the Terminal

The following commands must be run from the terminal. Information on accessing via the terminal can be found at the Accessing HPC page. 

Preinstallation Warning

If you have initialized Conda in your base environment (your prompt on Greene may show something like (base) [NETID@log-1 ~]$) then you must first comment out or remove this portion of your ~/.bashrc file:

The above code automatically makes your environment look for the default shared installation of Conda on the cluster and will sabotage  any attempts to install packages to a Singularity environment. Once removed or commented out, log out and back into the cluster for a fresh environment.

Prepare Overlay File 

Above we used the overlay file "overlay-15GB-500K.ext3.gz" which will contain all of the installed packages. There are more optional overlay files. You can find instructions on the following pages: Singularity with Miniconda, Squash File System and Singularity . 

Launch Singularity Environment for Installation

Above we used the Singularity OS image "cuda12.3.2-cudnn9.0.0-ubuntu-22.04.4.sif " which provides the base operating system environment for the conda environment. There are other Singularity OS images available at /scratch/work/public/singularity

Launching Singularity with the --overlay flag mounts the overlay file to a new directory: /ext3 - you will notice that when not using Singularity /ext3 is not available. Be sure that you have the Singularity prompt (Singularity>) and that /ext3 is available before the next step:

Install Miniforge to Overlay File

Next, create a wrapper script at /ext3/env.sh

Your /ext3/env.sh file should now contain  the following:

The wrapper script will activate your conda environment, to which you will be installing your packages and dependencies. 

Next, activate your conda environment with the following:

Install Packages to Miniforge Environment

Now that your environment is activated, you can update and install packages

To confirm that your environment is appropriately referencing your Miniforge installation, try out the following:

Now use either conda install or pip to install your required python packages to the Miniforge environment.

To install larger packages, like Tensorflow, you must first start an interactive job with adequate compute and memory resources to install packages. The login nodes restrict memory to 2GB per user, which may cause some large packages to crash.

After it is running, you’ll be redirected to a compute node. From there, run singularity to setup on conda environment, same as you were doing on login node.

Configure iPython Kernels

To create a kernel named my_env copy the template files to your home directory.

To set the conda environment, edit the file named 'python' in /.local/share/jupyter/kernels/my_env/. 

The python file is a wrapper script that the Jupyter notebook will use to launch your Singularity container and attach it to the notebook.

At the bottom of the file we have the template singularity command.

WARNING: If you used a different overlay (/scratch/$USER/my_env/overlay-15GB-500K.ext3 shown above) or .sif file (/scratch/work/public/singularity/cuda12.3.2-cudnn9.0.0-ubuntu-22.04.4.sif shown above), you MUST change those lines in the command above to the files you used.

Edit the default kernel.json file by setting PYTHON_LOCATION and KERNEL_DISPLAY_NAME using a text editor like nano/vim.

to

Update the "<Your NetID>" to your own NetID without the "<>" symbols.

Launch an Open OnDemand Jupyter Notebook

https://ood.hpc.nyu.edu

Configure and Launch your Notebook

Select kernel

Once configured and launched, kernels can be selected in the "New" dropdown or within the notebook under the kernel menu. Please note that your notebook view may look slightly different depending on available directories and environments, as well as if you choose the lab or traditional notebook view.