FAQs - statonlab/UTIA_Computational_Resource GitHub Wiki

We are excited to announce that new high performance computational equipment to support high throughput, intensive data analysis is now available to all UTIA students, staff and faculty. Two new Linux servers with data storage are online and ready for use. This equipment is suitable for multi-terabyte genomics, proteomics, metabolomics and phenomics projects, including tasks such as genome assembly, read/spectra mapping, variant calling, association analysis, image processing, machine learning/deep learning, and more.

We thank UTIA AgResearch and the UTIA Genomics Center for the Advancement of Agriculture for their generous support in funding the equipment, and Dr. Meg Staton’s lab, the Genomics Hub, and OIT for setting up and managing the equipment.

What are the machine specs?

We have two servers and a storage array.

Sphinx (Server, Designed for I/O intensive jobs):

  • Short Answer: 64 CPUs, 512Gb RAM, 70Tb SSD
  • Details:
    • Dell PowerEdge R7525
    • Red Hat Enterprise Linux 8.4 (Ootpa)
    • AMD EPYC 7302 3.35GHz 16-Core Processor x 2 = 32 cores = 64 logical CPUs
    • 512 Gb RAM (DIMM DDR4 3200 MT/s)
    • 70 Tb of solid state drives

Centaur (Server, Designed for multithreaded or parallel jobs):

  • Short Answer: 96 CPUs, 512Gb RAM
  • Details:
    • Dell PowerEdge R7525
    • Red Hat Enterprise Linux 8.4 (Ootpa)
    • AMD EPYC 7402 3.35GHz 24-Core Processor x 2 = 48 cores = 96 logical CPUs
    • 512 Gb RAM (DIMM DDR4 3200 MT/s)

Pickett (Storage Array):

  • Short Answer: 435Tb storage, RAID backups
  • Details:
    • Dell EMC ME484 Storage
    • 235Tb storage mounted on both servers (NFS
    • 200Tb storage mounted on centaur (iscsi)
    • 12TB 7.2K RPM NLSAS 12Gbps 512e 3.5in Hot-plug Hard Drives
    • Double parity RAID (6) configuration

What can I do with this equipment?

We welcome all users from UTIA for whatever data intensive research is needed, which might include but is not limited to multi-terabyte genomics, proteomics, metabolomics and phenomics projects, including tasks such as genome assembly, read/spectra mapping, variant calling, association analysis, image processing, machine learning/deep learning, and more.

Users are welcome to install software locally (no root access provided), and if that is not possible, we can also provide software installation help to support your specific needs.

How is this different from ISAAC (formerly the ACF)?

This equipment is designed to be more flexible and easier to use than the high performance cluster managed by OIT. The primary difference is that it will not require use of a job scheduler submission system such as SLURM or Open PBS. This provides direct benefits to our users: no complex scripting requirements or strict job time limits. Users will be able to directly ssh into each server and run jobs from the command line. However, because we are not using a software scheduler, sharing of the equipment will have to be managed through a centralized group (the Staton lab) and communication among users and use of a calendar are essential and will be enforced.

Further differences include:

  • we are not planning to charge for computational usage or storage at this point in time; please note, we are not a long term repository for data storage (more info below)
  • this equipment is managed and owned by UTIA, with OIT as a contracted partner for physical housing and systems administration support

Where is it located?

The equipment is housed in the Kingston Pike Building server room with other UTK computers. The 24/7/365 monitored environment has redundant power, network, cooling and physical security. All access to the equipment will be via ssh, i.e. online only.

How do I start using it?

New users are required to attend an approximately 1 hour orientation to learn about the equipment, including how to get access, rules for usage, coordinating usage with others, data storage policies, etc. To sign up, please contact Meg Staton (Associate Professor, EPP, [email protected]) and Matt Huff (Research Associate, EPP, [email protected]).

Where can I get help?

Currently the Staton lab is managing the equipment, led by Meg Staton (Associate Professor, EPP, [email protected]) and Matt Huff (Research Associate, EPP, [email protected]). All inquiries and communications about the equipment should be directed to them. New users are required to attend an orientation, see above.

Check the wiki home page for office hours.

Is my data secure on this system? Can I store it there forever?

No. Our storage system is backed up by RAID only, which means if a hard drive fails, your data will NOT be lost. However, there are NO off site backups, and this equipment is NOT a long term storage solution. We strongly encourage all researchers to transfer raw data and final analysis results to other long term storage solutions, including Google Drive, which currently offers unlimited cloud storage for through your UT system user account. Suggested options and transfer instructions will be covered in orientation and are also well documented in our user guides. We are happy to meet with anyone who needs additional help with this, and OIT also offers other data storage options and recommendations.

Who is managing the equipment?

Currently the Staton lab is managing the equipment, led by Meg Staton (Associate Professor, EPP, [email protected]) and Matt Huff (Research Associate, EPP, [email protected]). All inquiries and communications about the equipment should be directed to them. We also have secured support from OIT for basic systems administration, which will require occasional maintenance windows with equipment downtime.