Research Computing at Michigan Engineering – DRAFT

About Us

The Research Computing Group at CAEN provides assistance to the Michigan Engineering community with high-performance computing resources to help accomplish research and teaching goals. CAEN, in collaboration with U-M Advanced Research Computing (ARC), is responsible for supporting these advanced computing resources.

High-Performance Computing

Several computing resources are available, and outlined below. To learn more, click on a link below or contact us at [email protected].

Campus Cluster Resources

  • GreatLakes – U-M’s campus Linux-based high-performance computing cluster, Great Lakes, supports multi-core and long-running computations.
    • This cluster is typically used for general research and teaching.
  • Armis2 – The U-M cluster for research that uses sensitive data.
    • This cluster is aligned with HIPAA privacy standards.
  • Lighthouse – U-M’s Lighthouse cluster offers space available when a grant requires a hardware purchase.
    • ARC will maintain the hardware for a maintenance fee, which includes access to some software.

Off-Campus Cluster Resources

Funding

  • UMRCP – The University of Michigan Research Computing Package provides the following at no charge on the clusters at U-M:
    • 80,000 CPU hours for use on Great Lakes, Armis2, or a combination of both
    • 10TB of high-performance storage for use on the cluster or within your lab
    • 100TB of archival storage

Note: PIs must apply for this account for their researchers by visiting portal.arc.umich.edu.

  • Cost-Share – If you need more computational time on the cluster, the College of Engineering can help pay for it, if you have an externally funded research gift/grant. 

Bring Your Own Server

Do you have a server you need housed for you?   There are a couple of options available on campus to host server hardware. 

Storage

We recommend the storage services listed below. (If you need other storage solutions or aren’t sure which is best for your needs, use the Storage Finder Tool to find out what options are available.)

Turbo* – High-capacity, reliable, secure, and fast storage solution.

  • Description
    • Tuned for large files but capable of efficiently storing small files. 
    • Supports storing of sensitive data on HIPAA compliant systems (e.g., Armis2 cluster). 
    • Part of U-M Research Computing Package (UMRCP) providing storage allocations to researchers.
  • Usage
    • General-purpose research data storage, including sensitive data.
  • Accessibility
    • Accessible via local computers, lab/office setups, and HPC clusters (Great Lakes, Lighthouse, Armis2).

Locker – Cost-optimized, high-capacity, large-file storage service for research data. 

  • Description
    • Supports large files and can be used with Globus for data sharing with external collaborators.
  • Usage
    • Designed for data that does not require frequent access. 
    • Best suited for data at the end of its life cycle.
  • Accessibility
    • Globus for data sharing, compatible with U-M HPC clusters.

DataDen*- Low-cost, highly durable storage system.

  • Description
    • Meant for preserving electronic data generated from research activities.
    • Disk-caching, tape-backed archive optimized for long-term storage of data not regularly accessed (weeks to years).
    • Suitable for grant requirements, particularly NIH and NSF.
    • Part of U-M Research Computing Package (UMRCP) providing storage allocations to researchers. Most researchers may not incur additional costs.
  • Usage
    • Long-term archival storage for data not regularly accessed.
  • Accessibility
    • Long-term archival storage for data not regularly accessed.

* included with the UMRCP

Regulated Data

Doing research with CUI or otherwise sensitive data? U-M has options for securely working with that data.

  • Armis2 – HPC cluster for working with sensitive data
  • Secure Enclave Services  – Virtual server environments that can be setup, so researchers can have a location to securely work with data up to ‘high’ classification. 

Getting Started

ARC offers training and workshops as well as lecture captures of previous workshops so you can learn to use the clusters.

We have documentation on an Intro to Linux, the Cluster, and we are available to teach these to a course or research lab. Email [email protected] to schedule an appointment if you’re interested. 

Frequently Asked Questions

Am I eligible to use high-performance computing? 

Michigan Engineering provides access to a Great Lakes resource account that can be used by any Engineering student, faculty, or staff member. Restrictions are in place, depending on use case. Read more…

How do I request an account to access Michigan high-performance computing clusters (Great Lakes, Armis2, or Lighthouse)?

To request user access to one of the HPC clusters, submit the ITS-ARC Login form.

Can I access the cluster through a GUI web-interface?

Yes, Open OnDemand enables users to access their HPC resources for Great Lakes and Lighthouse clusters. 

Where can I find training on how to use the cluster? 

Visit our Intro to HPC page to find a variety of helpful documentation and other resources for new users. In addition, ARC provides several  training videos and periodically offers events/workshops

What is the difference between a user account, a SLURM account, and a paid/UMRCP account?

The user account enables the use of your UMICH uniqname and password to be used to log you into the computing clusters. Without requesting a user login, you will be unable to login to the cluster.

The SLURM account and paid/UMRCP accounts are resource accounts that let you use the resources on the cluster. They keep track of what resources you have used and bill your research group appropriately.

How do I move data in/out of the cluster/storage volume or share externally with Globus?

Globus is a data transfer tool that allows you to schedule transfers asynchronously and can resume transfers if it is interrupted. It can also be used to share data with external collaborators and can be used to schedule transfers at certain intervals so it can act as a primitive data backup tool.

I’m part of a student team or organization. Can I get time on the cluster? 

Sponsored Student Organizations can get time on the cluster at no cost. Ask your sponsoring administrator to submit a request for time on the cluster on your behalf.

How can I check if my shortcode qualifies for a cost-shared account?

Enter your shortcode into the chartfield converter. If the “Fund” entry has a 20000, 25000, or 30000, you qualify for a cost-share.

How can I request a change to my Turbo volume or start a new one?

Turbo is a high-capacity, network storage solution available to U-M users. You can change or order additional Turbo storage via the SRS website.

I’m writing a grant. How do I include info on the cluster in my proposal?

ARC has provided information about the cluster that you might need with submitting a research proposal.  If you would like to give an acknowledgment for the cluster, ARC has prepared some language for that.

I’m writing a grant. How do I purchase hardware as part of my grant?  

For researchers with a requirement to purchase hardware, Lighthouse supports the addition of hardware in an environment similar to Great Lakes. You are encouraged to contact us to discuss your purchase. For more information, contact [email protected].


Contact Us

Contact [email protected] with questions or concerns about high-performance research computing at Michigan Engineering.