RCC 2024 Tactical Objectives Retrospective

Thursday 12/12/2024

With 2025 just over the horizon, the Research Computing Center (RCC) would like to reflect on the past year’s objectives.

 

Finished migration of Research Archival volumes to the new Ceph platform 

The RCC completed the final data transfer for all Archival customers to our new Ceph-based filesystem in June of 2024.  Additionally, we were able to merge our two Archival storage servers into one unified system. The 2.5 petabyte Ceph filesystem provides better stability and manageability and makes it easier to increase storage on the system.

 

Added more GPUs to the High-Performance Computing (HPC) cluster 

The RCC continues to provide high-demand GPU resources to the FSU research community. In 2024, 12 GPU servers were added with 24 NVIDIA A4500 cards (2x per system) to the cluster, with plans to expand to more in 2025.

 

Implemented liquid cooling in the Sliger Datacenter 

In Q2 2024, after substantial upgrades to the infrastructure at the Sliger Data Center, we deployed 28 liquid-cooled compute nodes to the HPC cluster. These were made available automatically to research groups that have bought into the HPC. We will be adding 12 more liquid-cooled nodes in Q1 2025.

 

Updated software packages 

Throughout 2024, the RCC added and upgraded fifteen software packages in the HPC cluster. Some of the more salient packages included:

  • Anaconda/Python 3.12 
  • PyCharm IDE
  • MATLAB 2024b
  • STATA 18
  • DFTB+(new)
  • RELION (new)

Check out the full software catalog to see versions of all packages.

 

Upgraded Open OnDemand to Version 3 

In February 2024, the RCC upgraded its HPC web interface, Open OnDemand, to version 3. This upgrade brings bug fixes and user interface enhancements to the portal. With Open OnDemand, you can upload and download files, create, edit, submit and monitor jobs, run GUI applications and connect via SSH. All this can be done via a web browser, with no client software to install and configure. Next time you are on the FSUSecure wi-fi or the FSU VPM we encourage you to give the Open OnDemand program a try.

 

Ongoing improvements to the self-service portal 

This past year, the RCC has made a number of improvements to our users in our self-service web portal:

  • Implemented CAS-based single sign-on login for all users
  • Added technical details about our cluster, including compute node details for each queue
  • Added purchase information for resource owners on the website
  • Introduced monthly emails to resource owners with a detailed summary of usage for both queues and storage volumes

 

Continued our Outreach and Education initiatives 

Part of the core mission of the RCC is to provide education and training opportunities to the FSU research community. In 2024, the RCC presented nine workshops with 225 attendees.

In addition, we debuted our HPC video training course, “HPC Driver’s Ed”. This course was designed to help new users understand the basics of HPC and how to best utilize HPC for research and instruction.

 

Additional items Accomplished

  • Improvements were made to the RCCTool, a command-line utility to see information about your HPC account, view usage details, and more.
  • Improvements to the RCC internal network configuration, removing bottlenecks, and bringing speeds up to 100Gbps for most interconnects.