ICE Scheduler, Software, and Storage Upgrades, May 2023

Since 2018, PACE has provided the Instructional Cluster Environment (ICE) to support educational efforts in scientific computing. ICE supports credit-bearing courses from across Georgia Tech as well as non-credit workshops.

In May 2023, PACE significantly upgraded and restructured ICE to improve the experience for students with more reliability, more flexible resources, and improved storage.  PACE has worked with summer instructors to prepare for the new environment, and instructors planning fall courses on ICE are encouraged to contact PACE for assistance in updating course materials throughout the summer. Detailed documentation on use of the new ICE environment is available to students.

ICE moved to the Slurm scheduler, a process already completed on PACE’s Phoenix and Hive research clusters. This aligned ICE with the research clusters and provides a more reliable scheduler. PACE also update all software on ICE, again matching the new environment on the research clusters.

The College of Computing and PACE collaborated to purchase a Lustre parallel filesystem, much like the one on the Phoenix research cluster, to provide new faster storage for ICE. This device provides faster performance than the home directories on ICE and is set up as scratch space for temporary storage.   

In another collaboration with CoC, PACE merged PACE-ICE and COC-ICE to form the new ICE. Overflow between nodes increases availability of resources for all courses and decreases wait times while preserving priority for CoC courses on former COC-ICE nodes and for other colleges’ courses on former PACE-ICE nodes. Students will also have the opportunity to experiment with the additional GPU architectures added to COC-ICE in the past year, including Nvidia A40 and A100 and AMD MI210 architectures.   

With a new scheduler, software environment, and storage, the merged ICE cluster provides an enhanced experience for Georgia Tech students building skills in scientific computing.