Zurada User Documentation
Policies
Zurada’s System Use Policies
User Account
Running Jobs
Data Storage (Disk Usage)
Large-Memory Node Utilization
GPU Resources Utilization
Installing packages system-wide
Accounts and Support
Request an account
Accounts for UofL individuals
UofL VPN Connection
Request Support (Tickets)
Getting Started
Usage Agreement
HPC system overview
About the cluster
About Scientific Software
About Jobs
Quickstart
Logging into the cluster
Using the command line
Using MobaXterm
Copying files to/from the cluster
Using the command line
Using MobaXterm
Using software installed in the cluster
List available software
Load software
List currently loaded software
Unloading software
Queues and jobs
Resource Restrictions
Running applications on the login nodes
Job runtime restrictions
Guides
HPC System Guide
Introduction
What is a Shell?
What is case-sensitivity?
Connecting to the Cluster
Understanding Filesystems
What is a Filesystem?
Filesystem Hierarchies
Navigating the filesystem
Using Modules
Slurm Queueing System Guide
Basic Slurm Terminology
Basic Slurm Terminology
Job Types and Job Submission
Interactive Jobs
Batch Jobs
Job Arrays
Job Dependencies
Job Environment
Slurm environmental variables
Strategies When Submitting Slurm Jobs
Single Job with Multiple Job Steps
Job Arrays
Requesting CPU Nodes with Multiple Parallel Tasks
Single GPU node with one task per GPU and CPU cores evenly distributed across tasks
Preemption
Preemptable Jobs
Storage Guide
Understanding Storage on Compute Nodes
Storage types based on node accessibility
Filesystem locations users should understand
Recommended Workflow
Copying Data Between Home and Scratch
Software
Conda (Anaconda/Miniconda/Miniforge)
Basics
About Anaconda, Miniconda and Miniforge
What is a Conda environment?
Why use multiple Conda environments?
Using Conda
Loading the miniforge3 module and the base environment
Creating and activating an environment
Installing packages with conda and mamba
Installing packages with pip
Cloning an environment
Miscellaneous
Conda in a batch job
Conda in an interactive job
Gaussian
Running Gaussian
Example Slurm Job Script
Jupyter
Launching Jupyter through a batch job
1. (Optional) Create a jupyter environment
2. Create the submission script
3. Connect to jupyter from your web browser
Launching Jupyter through an interactive job
1. (Optional) Create a jupyter environment
2. Submit an interactive job
3. Manually launch Jupyter
4. Access Jupyter from your workstation
Transitioning from Jupyter to Python Script
1. Export the Notebook
2. Clean Up the Script
3. Replace Notebook-Specific Features
4. Handle File Paths and Inputs
5. Test the Script
MATLAB
Basics
Workers and pools
Parallel and distributed execution
Cluster profiles
Submitting a batch job
Submit jobs through a batch script
Submit jobs through MATLAB’s command prompt
Submit jobs through a batch script and a MATLAB submission script
Creating a cluster profile for Zurada
LAMMPS
Running LAMMPS
Example Slurm Job Script
Building LAMMPS
PyTorch
Verifying GPU Availability
Using GPUs in PyTorch
Moving Tensors to GPU
Model Training on GPU
Monitoring GPU Usage
Multi-GPU Usage in PyTorch
Single Node, Multi-GPU (DataParallel or DDP)
R
Using R
Installing R Packages
Installing R packages in custom locations
Installing R Packages with External Library Dependencies
Example: Installing the
units
Package
Example: Installing the
sf
Package
Simplifying with Conda
Using the
pak
Package Manager
RStudio
Pre-launch
Launch RStudio Server
Tensorflow
Verifying GPU Availability
Single Node, Multi-GPU Training
VASP
License Restrictions
Running VASP
VASP on GPU nodes
VASP on CPU-only nodes
Example Slurm Job Script
AI Use Cases
Pneumonia detection based on Chest X-Ray
1. Ingesting the data
2. Training the models
CNN Training and Validation
Transfer Learning Training and Validation
Fine Tuning Training and Validation
3. Visualize Metrics
4. Save your results for further analyses
Med-BERT
Zurada User Documentation
Guides
View page source
Guides
HPC System Guide
Introduction
What is a Shell?
What is case-sensitivity?
Connecting to the Cluster
Understanding Filesystems
What is a Filesystem?
Filesystem Hierarchies
Navigating the filesystem
Using Modules
Slurm Queueing System Guide
Basic Slurm Terminology
Basic Slurm Terminology
Job Types and Job Submission
Interactive Jobs
Batch Jobs
Job Arrays
Job Dependencies
Job Environment
Slurm environmental variables
Strategies When Submitting Slurm Jobs
Single Job with Multiple Job Steps
Job Arrays
Requesting CPU Nodes with Multiple Parallel Tasks
Single GPU node with one task per GPU and CPU cores evenly distributed across tasks
Preemption
Preemptable Jobs
Storage Guide
Understanding Storage on Compute Nodes
Storage types based on node accessibility
Filesystem locations users should understand
Recommended Workflow
Copying Data Between Home and Scratch