Hydro User Documentation

Introduction

The Hydro HPC system provides highly productive resources and services for computational and data-intensive science. It is designed for maximum throughput on large-scale, complex applications using InfiniBand interconnect architecture. Hydro consists of 70 server nodes with a mix of AMD and Intel CPUs, and 18 NVIDIA A100 GPUs, 6 PB of online storage and 430 Gbps in external networking capability. The architectural balance makes Hydro capable of supporting an extremely wide range of science disciplines and uses, from at scale simulation and modeling to the most advanced approaches to Big Data and Artificial Intelligence. Hydro supports a current Linux base software stack and programming environment.

Getting Started with High-Performance Computing

There are no specific prerequisite courses or high-performance computing (HPC) experience required before using Hydro. However, if you are unfamiliar with using an HPC cluster, it is highly recommended that you take NCSA’s short tutorial Using an HPC Cluster for Scientific Applications before continuing.

Browse NCSA’s HPC-Moodle for a full list of HPC training opportunities including self-paced tutorials and training events (in-person and virtual).

---

User Guide

• System Architecture
  • Hardware Description
  • Storage and File Systems
• Illinois Computes
  • Illinois Computes Participation
  • Illinois Computes Specifics
• Accessing Hydro
  • Access and Policy
  • Logging In
  • Account Management
• File Management
  • Transferring a Few Small Files
  • Transferring Many Files or Large Files with Globus
• Programming Environments
  • Software
  • Shells and Modules
  • Programming Environments
  • Compiling
  • Other Programming Environments
  • Python
  • Anaconda
  • anaconda3_cpu
  • anaconda3_gpu (for CUDA)
  • Installing packages
• Containers
  • Using Docker Images with Apptainer
  • Interacting with Host Filesystems
  • Mounting Images of Many-File Datasets
  • Running with GPU Acceleration
  • Running on Multiple Nodes with MPI
  • Accessing Hydro Modules in a Container
• Partitions and Job Policies
  • Hydro Partitions (Queues)
  • Project and Job Accounting
• Running Jobs
  • Running Batch Jobs (Slurm)
  • Jupyter Notebooks
• File and Directory Access Control List (ACL)
  • Home Directories
  • Project Directory
  • Enabling ACL Modification (if needed)
  • See Also