Case Study: University of Illinois at Chicago achieves flexible, scalable research computing with Liqid composable infrastructure

UIC Utilizes Liqid Composable Infrastructure for Uneven Applications in Scientific Research

The University of Illinois at Chicago’s Electronic Visualization Laboratory (EVL) needed a more flexible way to support diverse, resource-intensive research workloads spanning deep learning, visualization, AI, and scientific computing. Traditional fixed cluster designs were too rigid and often left expensive GPUs and other accelerators underutilized, making it hard for researchers to scale efficiently or match infrastructure to each application’s needs. Liqid’s composable infrastructure was selected to address these challenges.

Using Liqid composable infrastructure, EVL built a pooled, software-defined environment with 64 GPUs, 24 CPU nodes, NVMe storage, 100G networking, and large memory pools, managed alongside Kubernetes for rapid, on-demand deployment. The result was a more agile platform that could assign the right resources to each workload, improving utilization and enabling reproducible experiments, peer-to-peer GPU communication, and faster scaling for data-intensive research. Liqid’s system is now supporting computational fluid dynamics, brain imaging, and high-resolution visualization at UIC.

University of Illinois at Chicago

Luc Renambot

Research Associate Professor