COMPaaS DLV

Composable Platform as a Service Instrument for Deep Learning & Visualization

SENSEI Panama, an application of Deep Learning and Visualization that lets anthropologists walk inside their data in virtual reality. Photo courtesy Lance Long, EVL.

In September 2018 the National Science Foundation awarded EVL a three-year grant (October 1, 2018 to September 30, 2021) to acquire the Composable Platform as a Service Instrument for Deep Learning & Visualization (COMPaaS DLV), with Maxine Brown as PI and Andrew Johnson, Robert Kenyon and G. Elisabeta Marai as co-PIs. NSF provided $997,363 for hardware toward a total project cost of $1,424,806, with UIC cost share covering staff and student support. COMPaaS DLV is a flexible, scalable, high-throughput composable infrastructure that treats CPUs, GPUs and storage as pools of resources provisioned on the fly for each application. It was initially made available to 17 faculty from five departments in the UIC College of Engineering for fundamental research in deep learning, visualization, and their combination, and connected through UIC’s research network to campus, regional and national computing facilities.

Original news post: COMPaaS DLV: Composable Platform as a Service Instrument for Deep Learning & Visualization