This summer and fall, the three-person University Research Computing (URC) group will phase-in significant upgrades to the education and research clusters of the High Performance Computing (HPC) environment.
UNC Charlotte faculty and researchers leverage the HPC to compute, store, and access incredibly large sets of data. Last year, Drs. Min Shin and Srinivas Akella from the College of Computing and Informatics computed complex algorithms with the HPC as principal investigators for a defense agency project. A standard computer could not have processed the 20 hours of video (367 GB) in the required five-day turnaround. That’s partly because the HPC includes powerful GPU cores, which increase processing power for images; and compute cores, each of which can operate and store memory separately.
From June to October, the URC team will double GPU cores to 160 and more than double compute cores from 1,328 to 3,092. Additionally, the team will expand IB, technology that supports simultaneous network connections, and will upgrade Hadoop, a software library framework that distributes processing of large data sets. Finally, the team plans to install a new core network for the teaching cluster.
Overall, HPC users can expect to experience significantly better performance following 2020 updates.