HPC upgrades increase speed and storage capabilities

Categories:News, URC

One of the keys to accelerating scientific, engineering, and business discoveries is the ability to analyze massive datasets and perform complex simulations that could paralyze even the most powerful personal computers. Enhanced computational resources now allow scientists, engineers, and researchers to overcome long-standing challenges, from those in the studies of climate change impact and genomes to biomedicine and beyond.

Upgrades to UNC Charlotte’s high performance computing environment during spring 2021 are expected to help facilitate the groundbreaking work of today and tomorrow.

data center stock

The updates equipped the environment with 40 new compute nodes, for a total of more than 7,000 compute cores on which work can operate. Compute nodes provide short-term storage, networking, memory, and the capability to run virtualized containers in the future using Singularity.

The research Hadoop cluster, which stores and processes large datasets, was upgraded with more powerful servers for an increase in storage—from 290TB to 1400TB, enabling researchers to analyze much more information more quickly than ever. The cluster provides a distributed file system for faculty and graduate student researchers, now with double the nodes (16) and almost triple the core count (320).

Soon, the University Research Computing team plans to incorporate a new GPU system with 8 NVIDIA A100 Tensor Core GPUs. The systems’ GPUs can exchange information at lightning speed to allow much more complex artificial intelligence and image processing than previous systems. University Research Computing also expects to model an emerging frontrunner of performance processors, the high-performing AMD EPYC processor with a new ten-node cluster. Faculty and researchers can then leverage both Intel and AMD architectures in their work.

Faculty, graduate student researchers and others can request access to the University’s high performance computing environment. To learn more, visit urc.uncc.edu.