The University of Queensland is upgrading its Bunya supercomputer by adding new GPUs and thousands of CPU cores. The project, known as “Bunya Phase 2.0”, aims to enhance the supercomputer’s capabilities in accelerated computing, machine learning, artificial intelligence, and interactive visual computing. The university intends to support a wide range of research workloads from various disciplines, including humanities, economics, molecular dynamics, genomics, weather and climate modeling, artificial intelligence, and astrophysics.
The upgraded system will feature Nvidia H100 and L40 GPUs, and AMD Genoa CPUs, with 200Gbps per node networking using Nvidia HDR InfiniBand smart edge switches. The university is also incorporating AMD Instinct MI210 GPUs to provide diverse options for different research use cases.
To ensure an “evergreen” approach, the university plans for iterative changes to the supercomputer rather than large one-off purchases, allowing them to stay up-to-date with the latest technologies. One important metric for evaluation will be “computational outcomes per watt of power consumed”, reflecting the growing concern for energy efficiency and environmental impact in high-performance computing.
The project is set to be completed by August or early September, providing researchers at the University of Queensland with enhanced computational power and capabilities for their research endeavors.
