The School of Computer Science at Waipapa Taumata Rau / The University of Auckland (UoA), has invested in and installed a new GPU supercomputer, aiming at building capability of large-scale machine learning and artificial intelligence research and teaching. 

The newly installed GPU supercomputer consists of two servers. Each server has eight Nvidia A100 80Gb SXM4 GPUs, dual AMD Epyc 7742 64C/128T CPUs, and 1TB of main memory. The servers use NVSwitch, the latest interconnect technology that enables extremely fast and scalable computation for large machine learning models. It allows running machine learning models at scale and multiple machine learning experiments in parallel. 

The investment joins similar machines recently installed at UoA to support AI research and will represent a step change in the School of Computer Science’s and The University of Auckland’s Machine Learning and AI research and teaching capability. 

The GPU supercomputer will support UoA machine learning researchers on their government-funded research projects, including the MBIE biosecurity technology spearhead project co-led by Dr. Katerina Taskova, the Marsden project of “Training natural language models to understand genomics and study gout in Māori and Pacific populations” co-led by Prof. Michael Witbrock, and the Marsden project of “Unsupervised spatio-temporal representation learning for dynamic graphs” led by Dr. Kaiqi Zhao. Researchers who work on challenging research projects with big data or deep learning will be able to produce high-quality research output at an unprecedented speed. The new machine will also provide the best ever support for early-career researchers to build a track record. 

As our first supercomputer that supports teaching and learning, the School of Computer Science will use this supercomputer for creating hands-on learning experiences in their postgraduate machine learning courses. With access to high-performance GPUs, students will learn to solve real-world machine learning problems. This creates new opportunities for The University of Auckland and Aotearoa to become a world leader in machine learning and artificial intelligence education.