ARCH Clusters Guide

Welcome! This site is the one-stop user guide for every high-performance cluster operated by Advanced Research Computing at Hopkins (ARCH)

The Advanced Research Computing at Hopkins (ARCH) is a shared computing facility at Johns Hopkins University dedicated to supporting research, discovery, and learning through the use and development of advanced computing technologies.

ARCH provides access to high-performance computing (HPC) resources, high-throughput data storage, advanced networking, and scientific support services to empower computational research across all disciplines.

Mission

ARCH’s mission is to enable research, scholarship, and innovation by offering:

• Access to high-performance computing and storage systems

• Reliable infrastructure for data-intensive computing and simulation

• Scientific consulting for research workflows, application optimization, and data strategy

• Education and training in computational tools and methodologies

Services and Support

ARCH offers a range of support services to enhance your computational research:

• Research planning and resource allocation strategy

• Traditional HPC simulations requiring large CPU, memory, and I/O

• Interactive**Parallel/scalable file systems for improved data throughput

• Workflow design and implementation for computational pipelines

• Scientific support for emerging domains such as machine learning and AI

• Application support including installation and tuning

• Short-term project data storage with file system management

• Container and virtualization support (e.g., Singularity)

• Advanced consulting in Python, Go, and other research programming languages

• Fast data transfer via Globus and JHU’s HORNET network

• Educational and training resources for students and faculty

• Colocation facilities for PI-owned hardware

History and Governance

ARCH was previously known as the Maryland Advanced Research Computing Center (MARCC).

• MARCC was a collaborative HPC initiative between Johns Hopkins University and the University of Maryland, College Park

• Located on JHU’s Bayview Campus, with offices at the Bloomberg Center on the Homewood Campus

• Funded by the State of Maryland through a grant to JHU via the Institute for Data Intensive Engineering and Science (IDIES)

More Information

Visit the official ARCH website for more:

📍 ARCH Website 📍 Rockfish Portal 📍 DSAI Portal 📍 Request Help 📍 Support

ARCH Systems

• Rockfish Cluster
  • Rockfish Quick Start
    • What you need
    • Create an Account
    • Log in
    • Load software modules
    • Quick interactive work
    • Submit your first batch job
    • Managing Python, Jupyter, RStudio
    • Storage locations
    • House-keeping
    • Need help?
  • Creating an Account on Rockfish
    • Creating a User Account
    • Creating a PI Account
    • Password Management
  • Resources
    • Rockfish Hardware Summary
    • Filesystems on Rockfish
  • Navigating
    • Logging in to Rockfish
    • File Transfers
  • Slurm on Rockfish
    • Example Batch Scripts
    • GPU Jobs
    • Viewing Job Status & Efficiency
    • Job Scheduling with Slurm
    • Available Partitions
    • How Job Priority is Calculated
    • SLURM Overview
    • Checking Resource Usage & Billing
    • Viewing Historical Usage and Efficiency
• DSAI Cluster
  • DSAI Quick Start
    • What you need
    • Log in
    • Submit your first batch job
    • Managing Python, Jupyter, RStudio
    • Storage locations
    • House-keeping
    • Need help?
  • Creating and Managing Allocations and Accounts
    • Requesting an Allocation
    • Requesting User Accounts
    • Storage Allocations
    • Password Management
  • Resources
    • DSAI Hardware Summary
    • Filesystems on DSAI
    • Available Software Modules
  • Navigating
    • Logging in to DSAI
    • File Transfers
  • Slurm on DSAI
    • SLURM Overview
    • Available Partitions
    • How Job Priority is Calculated
    • Viewing Job Status & Efficiency
    • Job Scheduling with Slurm
    • GPU Jobs
    • Example Batch Scripts
• EDU Cluster
  • EDU Cluster Quick Start
  • EDU Cluster Hardware Summary
  • Filesystems on the EDU Cluster

Common Tasks

• Bash Quick Start
  • Understanding Bash Startup Files
  • Protecting Your $PATH
  • Aliases: Quick Commands
  • Helpful Customizations
  • Example .bashrc Template
  • Testing Your Configuration
  • Copy Default Bash Files
  • Troubleshooting
  • Further Resources
• Manage Group Permissions
  • Overview
  • Checking File Permissions and Group Ownership
  • Changing Group Ownership
  • Viewing Your Group Membership
  • Setting the Active Group
  • Default Group on Login
  • Summary of Useful Commands
• Connecting to the JHU VPN
  • Step 1: Download Ivanti Pulse Secure
  • Step 2: Configure the VPN
  • Step 3: Authenticate
  • Step 4: Disconnect When Done
  • Troubleshooting
• GPU Utilization: Measuring, Diagnosing, and Improving
  • Measure GPU Utilization in Real Time
  • Reviewing GPU Job Statistics
  • How to Improve GPU Utilization
  • Zero GPU Utilization (0%)
  • Low GPU Utilization (< ~15–30%)
  • Common Mistakes
  • Build Your Skills
  • Getting Help

Tutorials

• Managing Software Using Lmod
• Containers
• Environments & Packages
• Interactive Interfaces
• Specialized Applications
• Workflows

Support and Help

• ARCH Terminology
  • Hardware Terminology
  • Software and Scheduling
  • Data and Storage
• Good Citizenship & Terms of Use
  • Definitions
  • Being a Good Citizen
• Citing ARCH
  • Acknowledging ARCH in Publications
  • Community Awareness and Impact
• Support
  • Open a Ticket
  • User Survey
  • Software and Feature Requests
  • Training & Workshops
• Frequently Asked Questions
• Service Level Expectations
  • Hours & Response Times
  • Where to Get Help
  • What We Expect from Users
  • What’s Out of Scope
  • Our Commitment