Virga: Australia’s AI Powerhouse

Australia has officially added another supercomputer to the TOP500 list with the implementation of Virga. Officially coming online in June 2024, Virga is the newest HPC system to come out of the Australian Commonwealth Scientific and Industrial Research Organisation (CSIRO.)

As of the June 2024 TOP500 rankings, Virga has achieved the 72nd position on the list. Built on Dell PowerEdge XE9640 servers and featuring Nvidia H100 Tensor Core GPU accelerators with 94GB of high-bandwidth memory as well as 4th Gen Intel Xeon Scalable processors, Virga was designed and built to optimize AI workflows.

Additionally, Virga features a Transformer Engine which will significantly speed up AI performance and capabilities and help train large models within days or even hours. The system also uses hybrid direct liquid cooling to reduce the need for energy-intensive air cooling.

In a statement announcing the machine, CSIRO’s Digital, National Facilities and Collections Executive Director Professor Elanor Huntington showed excitement for the machine’s future.

“AI is used in practically all fields of research at CSIRO, such as developing world-leading flexible printed solar panels, predicting fires, measuring wheat crops, and developing vaccines, just to name a few,” Huntington said. “High-performance computing systems like Virga also play an important role in CSIRO’s robotics and sensing work and are crucial to the recently launched National Robotics Strategy to drive competitiveness, and productivity of Australian industry.”

Housed at the CDC Hume Data Centre in Canberra, Virga is expected to replace CSIRO’s previous Dell-powered HPC system called Bracewell. While Dell was originally contracted to create the system in 2023 with an AU$14.5m (US$9.65m) bid in 2022, the system ended up costing AU$16.3m (US$10.85m) to build.

While this system will have a wide variety of uses, Jason Dowling from CSIRO’s Australian e-Health Research Centre stated in the press release that Virga will be used frequently to solve medical problems. Specifically, he believes the machines will allow researchers to train and validate new computational methods, thereby helping medical professionals develop translational software in medical image analysis for image classification, segmentation, reconstruction, registration, synthesis, and automated radiology.

More specifically, Dowling points out that Queensland Children’s Hospital will benefit greatly from this new machine. He states that this machine will help train AI models to diagnose pathology from MRI scans of the lungs of children with cystic fibrosis.

The launch of Virga marks a significant milestone in Australia’s journey toward implementing AI tools within its HPC ecosystem. As Virga begins to tackle complex challenges across various scientific disciplines, from climate modeling to medical imaging, its impact will likely extend far beyond the realm of pure research.